Home

User Manual - PIVTEC GmbH

1. detected for each interrogation area J ca a Step 2 The line segments are ae subsequently assembled into sets of aiai continuous lines spanning the entire to store gt UO PIVmap T01 x x a restore image Separation into two major config Interactive tool for determining directions as well as sorting of the lines O cel ae he are then performed to reconstruct a Sum er grid meshing In order to properly overlay the images the origin can be shifted in grid intervals using a desired DW ping Dewarping point of reference in the calibration OUTPUT Raa a ar grid Camera A Camera B Step 3 At this point the projection coefficients can be calculated using a desired magnification factor for the back projected image For more information about this module please consult the paper by Ehrenfried 2001 Some notes on obtaining calibration grid images The grid must be made up of two sets of continuous evenly spaced parallel lines that are orthogonal to each other The grid should either have bright lines on a dark background or dark black lines on a bright white background The latter lends it self especially well to be printed by a laser printer The grid should be aligned parallel with the light sheet of cover the entire field of view if possible The grid image should contain 10 30 lines in each dimension roughly 30 60 pixel spacing this results in 100 to 1
2. 8 PIVview User Manual mouse cursor position Zoom out Zoom in change the image magnification The scroll bars on the edges can be used to pan through the enlarged image data set Shortcut keys are and Brighter Darker changes the visual brightness of the currently loaded image The image data itself is not affected The shortcut keys are b and d Auto scale sets default intensity on image Change LUT changes the current image lookup table LUT or switches to previous LUT Clicking on the function button repeatedly cycles through all possible LUTs Currently there is no provision for user defined LUTs Shortcut key is L Clipping enables or disables image clipping for image intensties exceeding the LUT maximum intensity With clipping on intensities greater than the maximum are set to the maximum With clipped off the intensity wraps back to the bottom of the LUT Shortcut key is c Show LUT controls the display of the LUT bar on the left side of the image Shortcut key is CTRL L Show Image is useful to hide the image to better view the computed PIV data Shortcut key is I The PIV menu Calibration puts PlVview into calibration mode hiding the current PIV data set and displaying a moveable ruler and origin Within a separate dialog the magnification factor and reference point coordinates can be defined interactivley After closing this menu PivView reverts back to the interrog
3. Perform ensemble correlation Rather than calculating the displacement field for each image pair this method computes the averaged correlation planes from all images before detecting any correlation peaks This procedure is common in the analysis of uPIV applications and reduce the effects of underlying Brownian motion Load processing parameters from file When enabled the images will be processed using the parameters in the specified file Otherwise the evaluation uses the currently defined parameters Batch Processing Output Options A variety of output options may be defined in this menu Preferably the data from each processing image should be stored in the platform independent netCDF format This data can then be exported in any desired form using the same batch processing menu Batch PIY Processing 2 x PELLLLLEPTPPPLPPPPPLLPFLPRLPFLPELPERE Stub for generated statistics files seg M Store data for each processed image file IZ NetCDF Platform independent Compute Reynolds stresses Save deviations from mean MV Other output formats Tecplot TM ASC y T Displacement raw data output including IZ Magnitude J SNA data J Correlation coefficients P All correlation peaks M Velocity data output including 4 Magnitude T Yorticity Normal strain I Lambda2 T Z Coordinate Shear strain Enstrophy Differentials T Generate plot for each processed image file Export format BMP J
4. Maximum Displacement Difference This method computes the magnitude of the vector difference between the vector in question and its eight surrounding neighbors When this difference exceeds the specified value for more than 4 neighbors then the analysed vector is labelled to be an outlier Normalized Median Filter applies a median filter to each vector along with its eight neighbors A vector is an outlier if the difference to the median displacement value is greater than the given threshold Dynamic Mean Filter This filter works very effectively even in flows with discontinuities e g shocks However it requires a careful adjust of the coefficients to work reliably see Raffel et al 1998 p 153 for more information Minimum signal to noise ratio disables all displacement vectors which have a SNR less than the specified value Ratio Peak 1 to Peak 2 disables all displacement vectors for which the ratio between the strongest and second strongest correlation peak is less than the specified value Minimum correlation disables all displacement vectors for which the correlation peak value is less than the specified value Global Histogram restricts valid vectors to be within a certain range of displacements This rectangular or elliptical domain has to be set interactively within the PDF window using the mouse PIVview User Manual 21 Outlier Replacement Schemes Once outliers have been detected they can be o
5. Origin The origin is used as the reference point for the physical coordinate system of the data set It can be translated within the image by holding down the left mouse button Origin Offset Additional offset in physical space units if reference point is not at 0 0 0 in physical space Magnification Factor This field is used for the calculation of the magnification factor by using the pixel length of the ruler within the image numerator and the actual distance in physical units denominator Only the physical distance can be modified here The ruler length can only be changed by altering the ruler end points in the image The Apply Button can be used to immediately see the effect of the conversion factors on the data and grid PlVview User Manual 23 PIV Evaluation Setup Filters etc The menu tab allows the selection of various post processing options SB PIY Evaluation Setup Grid Correlation Algorithm Peak Detection Validation Conversion Filters etc Median kermel size a F Data smoothing Smooth kermel width in nodes i Differentiation Calculate cell size for differentiation User defined differentiation cell size Horiz 3 Nodes vert 3 Nodes Extrapolate to edges Apply Cancel Median Filtering The final data set is once median filtered using either a 3x3 or 5x5 kernel Data Smoothing The final data set is once smoothed using a Gaussian weighte
6. Put images data into separate destination directory gt Choose fe fdpiv tmp Cancel Stub for Statistics Files specifies the beginning letters of the statistics files generated during batch processing These data sets are computed over all available files and are stored after the batch processing is completed For the stub seq the output files will be segAvgPiv dat mean pixel displacement segAvgVel dat mean velocity seqRmsPiv dat RMS pixel displacement seqRmsVel dat RMS velocity PIVview User Manual 37 Store data for each processed image file When enabled this option will store the displacement and velocity data for each processed image or file Otherwise only the mean and RMS data sets computed from all images will be stored Generate plot for each processed image file when enabled this option will generate an image file plot for each processed image using the currently defined plot configuration This is useful if animated sequences of the data are to be compiled later on Note These files are quite large and may result in the use of much disk space Put image data into separate directory this option can be enabled to store the computed data in a directory different from the image directory This can be useful when batch processing images retrieved from a read only medium e g CD ROM Definition of statistics terms Statistical quantities such as the mean and standard deviation of velocity
7. The following ccettictents PA AL TE for for files information are required canes Camera B gt a set of simultaneous acquired PIV IN IN IN IN images for both camera views e g 1 PIV Maplmage Maplmage image pair per camera oda ned from PIVmap T01 from PIVmap T01 gt images of a calibration grid placed OUT OUT parallel to the light sheet one image per D dl Pai D dl Pai camera The grid should also have some et S clearly identifiable reference marks on it to identify its orientation gt the position of the cameras in physical space with respect to reference grid more on this later Standard 2 C PIV Processing IN DPIV UT O 2 C PIV Data File netCDF IN netCDF merges PIV data sets from each camera view OUT StereoGeom specifies the stereo viewing geometry The first step involves determining two sets of coefficients which allow the PIV recordings to be matched to each other This projection of the images into a common image plane shall be referred to as image back projection The grid images are essential for determining the projection coefficients as the spacing between the grid lines is well defined The modules PIVmap3 and PlVmapT01 formerly named T01 use semi automated line or dot grid detection algorithm for exactly this purpose Once the projection coefficients are known PIV images for each view can be back projected The c
8. Base angle U dda Mapping Iterations ho 3 Flipl A f 1st Order Flip T B Mix nd Oth C Mix nd Tat 2nd Order Horiz Shift shifts the vertical Y axis horizontally across the image Defines the horizontal coordinate of the image origin Vert Shift shift the horizontal X axis vertically across the image starting from the bottom Defines the vertical coordinate of the origin Flip L R flips the image left to right Useful in case of using mirrors or viewing the light sheet from two different sides Flip T B flip the image top to bottom Iterations maximum number of Iterations used for the calculation of the coefficients low value faster higher chance of result not being converged high value slower better chance of convergence 1st Order sets mapping function to be a ratio of polynomials of first order Mix 2nd Oth mapping function only a second order polynomial divisor is unity Mix 2nd 1th mapping function ratio of 2nd order polynomial and first order polynomial divisor 2nd Order mapping function ratio of two second order polynomials Note higher order mapping functions may result in more precision but may require more iteration steps 64 PIVmapT01 Dialog Projection PlVview User Manual This dialog is used to set the image size of the back projected images Calculate button calculates back projected images which are displayed in the main window o
9. Binarize converts the image to contain only black 0 or white 255 pixels by setting values above the specified threshold to 255 Lower values are set to 0 Low Pass Filter applies a Gaussian weighted smoothing filter to the image In combination with the high pass filter the resulting image will be band pass filtered Recommended kernel widths are the range of 0 5 1 5 pixels Scan Tab Feature extraction in P Vmap3 is based on a correlation approach in which a template of either a dot or cross is matched locally to the image The position of best match indicates the PlVview User Manual 51 center of the dot or cross in the locally sampled area The performance of this matching process depends on the choice of sample size sampling overlap and a few more parameters that are set in this tab Spot Template radius Pi map Control El specifies the radial dimension of the dot or cross in the Fiters Scan Scale Map template Template radius px 4 Sample size Spot covers 16 7 of sample provides a choice of image sampling areas Typically A gt this size should roughly correspond to the mean Sample size pa 48x48 E Spacing of the features to be extracted Sample overlap 80 Sample overlap Minimum count E specifies the image sampling distance A higher Merging radius px sampling rate increases the overall performance but en decreases Processing speed Reconstruction Tolerance fics Minimum count As eac
10. Coon T Dynamic Thresholding Histogram Clipping klinimum Bz Maximum 30 f High Pass Filter Kernel size BE W Use Gaussian Weighting Clip at Zero T Binarize Threshalding Threshold 5 W Anti lias Smoothing O Fixed 3x3 Gaussian fc Warable Gaussian Kernel px 0 5 aj Tab General Always Split PCO Images PIV image pairs recorded with the PCO acquisition software are generally stored in a common file The upper half of the image is associated with the first exposure while the lower half corresponds to the second exposure This operation splits the input file into the two images Otherwise the software will try to find the matching second image Rotation rotates the image s in 90 increments This operation as well as all others is performed on both images of a pair Mirror can flip the image either left to right or top to bottom Both flip operations together are equivalent to a 180 rotation Tab Filters Filters are applied in the order they appear in the menu Convert Images to Floating Point Data prevents truncation of data during filtering procedures may increase overall precision Invert computes the negative of the image by turning bright pixels into dark ones and vice versa Median Filter performs median filtering on the data using a 3x3 or 5x5 kernel Dynamic Thresholding performs percentage based thresholding of the image This operation determines the minim
11. Qin Er ee 13 PIV Evaluation Setup E eT Tele ee ee 14 Phase correlation Phase only correlation sse eee ee e eee eee e 14 MiCMISIG Corral OM T 14 PIV Evaluation Setup AlgOrtNM etii aa E 15 Advanced PIV Processing OPIONS au ae tan ae 17 PIV Evaluation Setup Peak Detection u002222000022000nennnnnnennnnnnennnnnnen nennen 18 PIV Evaluation Setup s QU casar se ea 20 PIV Evaluation Setup Conversion uuuus222suusnnnnnnnennnnnnnennnnnnnennnnnnnennnnnnnen nennen nnennnnnnnn 22 PIV Evaluation Setup Fillers8le use eh 23 REGION O FINE Murcia 24 Plot Configuration MG AU sasa aac aaa olas 25 Vector Piloting Paramete S ivi A AAA E 25 OUtlier CISDIAY eere tii tenis 25 OMS V COLON oi idas 25 HA PA hace S NEK eee aan 26 Veco Envia lo eee ee eee 26 GontourPIOHING Paramelerscn cinco tegen ohne oa 27 GONIOUr TY eT 27 Gontour Display Te 28 PlVview User Manual V AXIS PIOWING Rarame tal Serres aida 29 Relorence EITE ads 30 FONE COS ari 30 The Preferences Men ios ici 31 lA O A A RE ee 31 Correlation Plane WING O Wii A e nene 32 Global Histogram WINDOW seen 33 I D PDES and Peak LOCKING ini nee 34 Bathi PIV PROCES SUING HEH 35 Bate Processing OB OMS ec ciara teeter shod ld 36 Batch Processing OUUU OPINAS ass 36 Denton oO Stasi CS Terms er rn 37 PiVmask Image Mask Editor a 22ER 38 PIE MENU Eee E ee 38 VIEW MENU nee nee a Tette sek 38 EUA NIE U o ee ee genen 39 Stereo PIV Modules un ende 40 st
12. Vector Fonts Colors Type of Ref Vector f Displacement C Velocity C Physical Shift T Show Reference Paint Show Reference Vector has to show a reference vector in the upper left corner It is only displayed in the presence of a vector map or during interactive interrogation Type of Ref Vector three different forms of labeling the reference vector are possible Displacement chooses a image based reference scale in pixels Velocity labels the vector with the currently active velocity unit Phys Shift uses the physical shift of particle images Fonts Colors E Piv iew Plotting Parameters x Vechors Contours Axis Ret vector Fonts Colors Eont Font Sizes Axis Labels ji 3 pt Asis Title i E pt aj Color preferences Vector color coding Outer vector Replaced vec Valid vector Apply OF Cancel Font allows the selection of the font type and size used for axis and LUT labels Color Preferences Vector color coding defines the color for various plot items and data types e g outliers vectors PlVview User Manual 31 The Preferences Menu Pry iew Preferences EE Priy iew Preferences EE Recent Image List contains ls Files Recent File List contains ls Files Fly File Extensions for netCDF files e Pl Image Extensions HIV Image Pairs r e BE fw BMP PNG FP IMG M TIFF PEM e IMA JPEG T PG
13. are calculated by the following equations Quantity Formula Notes Mean value IX here N is the number of valid measurements x x E gt X including interpolated values i 0 Variance 2 ea o i 0 i 0 l N N 1 Standard deviation o defined as square root of the variance Mean velocity u v w calculated from sequence of PIV data sets components using formula for mean value given above RMS velocity 0 10 40 using formula for standard deviation 1 wv Values for v v w w and v w are calculated ror 2 i uu 2 u u correspondingly Py Only validated and interpolated vectors are a Y a Ga used in the calculations N 3 A multiplication with density p results in the respective Reynolds stresses Fluctuation terms 1 _ _ uw 20 u w W i 0 HOD 38 PlVview User Manual PIVmask Image Mask Editor The program PlVmask allows the creation of image masks that can be overlaid on PIV images to prevent processing in these areas Any PIV image can be opened using the File Open function Several simple shapes then are available for the creation of the mask Once finished the mask can be stored as a bitmap file To enable its use by PlVview the appropriate option in the image preprocessing menu must be activated PivMask C DPI p1 ima E oxi File View Draw Help zeri K N eln X _ Left mouse button to add vertices right button to Finish PIVmask in edit mode displaying
14. as output grid If the grid from view A does not coincide then this data is interpolated PIVview User Manual 73 Shortcut keys in PlVview F2 region of Interest selection if enabled F4 PIV parameter menu also with right mouse button on image F5 start PIV evaluation F6 show correlation plane F7 show 2 D scatter plot histogram F10 redraw current view B image brighter D image darker show hide image after evaluation L change pseudo color look up table Shift L show previous pseudo color look up table P open plotting parameter menu C show hide scalar image after evaluation V show hide vector field after evaluation increase image magnification zoom in decrease image magnification zoom out CTRL i image pre processing menu CTRL L hide show LUT bar CTRL M open mask editor CTRL O open file dialog for loading images CTRL P print currently displayed data or image CTRL Q quit PIVview space bar toggle between images of image pair CTRL Space overlay images of image pair Mouse ShortCuts Actions Left click on image pair single point PIV evaluation using current parameter settings Right click on image pair opens context sensitive menu Double click on evaluated image opens a data monitoring window showing statistics for each point currently captured by the mouse 74 PIVview User Manual Supported File Formats Supported Image Formats Tagged image file format TIFF at 8 bit
15. average lt datalN_1 gt lt dataIN_N gt Parameter description lt dataIN_1 gt input first file in sequences to be processed lt dataIN_N gt input last file in sequences to be processed Command line options prefix lt stub gt overrides the output file stub to the name provided rms also stores the RMS data in a separate file reyn computes Reynolds stresses from velocity data F wee computes fluctuations by removing mean from each data set SO provides feedback during processing of data tec store result in Tecplot ASCII file o store result in Tecplot binary file DPIV The program dpiv reads the specified image s and parameter file and then performs complete PIV processing including gt image pre processing filtering background subtraction etc gt correlation analysis particle image shift extraction gt data validation gt conversion to velocity or physical shift The result is stored in a netCDF file such that it can be read by other P Vview programs Use ncdf2tec to convert the data to a Tecplot readable format Unless specified with the o option the output file name is determined from the image name Calling procedure dpiv lt par_file gt lt imageA gt lt imageB gt Parameter description lt par_file gt input PlVview generated parameter or data file netCDF format with processing information 68 PIVview User Manual lt imageA gt input PIV or speckle image lt imageB gt input Second
16. can only be loaded if the mapping coefficients are available This prevents processing of images that have not been registered with respect to each other Stereo Menu Options amp Toolbar E New project allows a step by step creation of a new stereo project using the StereoWizard program including the image reconstruction back projection procedure E E E E Z Th F f l a ka h 42 PlVview User Manual Open project opens an existing stereo project It loads previously calculated projection coefficients the viewing geometry the calibration images used by PlVmap as well as previously loaded PIV images if available Save project saves the current stereo configuration coefficents geometry amp parameters in the current file name Save as allows the stereo configuration to be stored under a different name Recent projects allows the selection of one of the 4 most recently accessed stereo projects Geometry opens the stereo geometry definition window module StereoGeom It is here where the camera position with respect to the viewing area light sheet is specified Image registration opens the PIVmap module which is used for determining the back projection coefficients Show Image A displays the image as viewed from camera A if loaded If the data has been processed the image will be overlaid with the corresponding 2 C vector or contour map s Show Image B displays the image as viewed from came
17. components of velocity Show Contour Image has to be enabled to show the contour image If a data set is available i e image is evaluated then enabling this option will immediate display the contour plot Bi linear interpolation is used to fill in the area between the individual data points UE Pi iew Plotting Parameters Vechors Contours Axis Ref vector Fonts Colors IZ Show Contour Image Default Contour Type X velocity m s Pixel Units f Physical Units e g Velocity Display Range Min 2 May 2 Nice Range Ho data available MINI E Lora Rainbow y I Discrete Contours 20 I Reverse LUT e Wrap LUT FF Fill Background Contour Type Specifies which kind of data is plotted as a contour plot The contour display range can be adjusted below Vorticity plots the out of plane vorticity as estimated by finite differences Ov Ou O LY Oy Shear Strain plots the in plane shear strain as estimated by finite differences AL gt y Ox Normal Strain plots the sum of the in plane extensional strains n as estimated by finite differences 7 0u Ov 1 E yy TE P Displacement Magnitude plots the pixel displacement vector magnitude Velocity Magnitude plots the velocity vector magnitude Physical Displacement Magnitude plots the actual physical displacement of the particle images or speckle patterns 28 PIVview User Manual Correlation Coeff
18. feedback of the processing results while also providing the capability of quickly obtaining a hardcopy of the calculated data 2 PIVview User Manual PlVview Versions Two versions of P Vview are available a demo version as well as a licensed version The demo version downloadable from PIVTEC s internet server provides nearly all features of the licensed version except for data export and batch processing The demo version can be distributed freely and is especially useful for optimizing a PIV experiment or for demonstrating the principles of PIV evaluation Full use of all of PlVviews features is only possible by obtaining a license through PIVTEC GmbH Credits PlVview was conceived and is maintained by C Willert PIVTEC GmbH G ttingen Germany and developed in close cooperation with the PIV Group of the German Aerospace Center DLR in G ttingen PIVview User Manual 3 System Requirements PIVview should run on essentially any of today s Windows operating systems 32 bit Microsoft Windows tested on Windows versions 95 98 NT4 2000 XP Pentium class processor gt 32 MB of RAM depending on Windows version 256 color SVGA or better graphics display hard disk space program executables 5 MB example files 30 MB utility programs 4 MB If you merely want to try out P Vview you need not install It on your system The program can be run from the CD ROM by double clicking the PivView2CDemo exe file Alternat
19. formatted data file h option shows only variable and attribute definitions and their values help option shows all available options for ncdump NCGEN The NCGEN program reads the specified raw CDF data file and generates a netCDF binary data file The makes it possible to convert the ASC Il output obtained from ncdump back to netCDF The resulting data file is written to the screen unless a file name is specified with the O option Calling procedure ncgen lt filelN gt Parameter Description lt fileIN gt input ASC II formatted data file with variable and attribute definitions o lt file0UT gt option stores the resultant netCDF data in the specified output file Otherwise data is written to stdout screen help option shows all available options for ncdump POSTPROC This program first builds a list of files from the first and last file provided and then performs post processing i e validation conversion filtering etc on each PIV displacement data set The results are placed in new files whose names are determined from the common file stub of the input files It is useful when only post processing of previously calculated PIV displacement data is necessary Calling procedure postproc lt PARfile gt lt dataIN_1 gt lt dataIN_N gt Parameter description lt PARfile gt input first file in sequences to be processed lt dataIN_1 gt input first file in sequences to be processed lt dataIN_N gt inp
20. gt ae DUR bE piwC1_0009 516 Add All pivC1_0006 b16 pivC1_0007 b16 da CI 0010 b16 S Remove l Available Files the left column shows the files currently available in the specified directory Multiple files can be selected by holding down the CTRL or SHIFT keys just as in the Windows Explorer Using the Add button the file selection is transferred to the list of files to be processed Selected Files The right column shows the files currently chosen for batch processing In analogy to the left column multiple files can be selected by holding down the CTRL or SHIFT keys Using the Remove button the file selection is removed from the list Directory this field displays the currently selected data image directory lt can be changed using the Directory button on the left File Type this field acts as a file filter showing only file which match the chosen extensions File type association is defined through File gt Preferences gt File Type Filter 1 Selects files in data directory with this prefixed filter Allows selection of all odd or all even numbered files Filter 2 A file filter that may be freely modified Caution Without the asterisc no files may be displayed 36 PIVview User Manual Batch Processing Options T Batch PIY Processing T Perform ensemble correlation F Only perform image pre processing and store resulting images T Load processing parameters from file i Choose
21. image of pair if required Command line options def performs second order accurate image deformation PIV mg performs multigrid pyramid PIV auto perform autocorrelation processing only one image is required NO PIP avoids flipping or rotating the image important for stereo PIV processing noip avoids image pre processing options norotate avoids flipping or rotating the image important for stereo PIV processing info provides feedback during processing of data o lt fname gt stores output data in file lt name gt pco treats the input image as a PCO double exposure image speckle perform speckle shift processing where the first image is considered the reference unshifted image wx y specifies the interrogation window size Overrides the value specified in the parameter file sxly specifies the interrogation window step size Overrides the value specified in the parameter file bx y specifies the global image offset Overrides the value specified in mask lt fname gt uses specified mask image Useful for batch processing with varying masks Note mask image must have the same size as PIV images DPIV3C The program dpiv3c reads the specified stereo configuration and parameter files to perform complete stereo PIV processing using the given image set This processing including image pre processing filtering background subtraction etc image back projection using projection maps including disparity c
22. kernel size of filter for smoothing the Hough transform Can normally be set to 1 e g no filtering Window size 128 Window step ESNE Image desampling Scanning Scan size theta angular increment used during maximum search Default of 5 usually does not need to be altered Scan size radius radius increment used during maximum search Default of 5 usually does not need to be altered Window size size of the sampling window used in the Hough transform Its size should be larger than the typically distance between two grid lines This window size also defines the length of the individual line segments In case of strongly curved grid lines the window size should not be chosen to be too large which might impede the accuracy Window step sampling distance between Hough interrogation windows This distance should be smaller than the window size to ensure overlapping line segments Small value slower but high accuracy High value faster less precise Image desampling down samples the image to speed up processing at the price of slightly reduced precision especially useful during optimization of line detection parameters Invert turns loaded image into its negative The Hough line detection algorithm requires bright line data on a dark background Detect lines button starts the detection algorithm This process is the most lengthy step in the grid reconstruction process but can be aborted at any point availa
23. normal to the light sheet Use the stereo configuration tool below to interactively position the cameras Camera Positions Camera A pe K a Camera B Ban b oo Origin a De Create new Stereo Pl Project Finally store the previous settings in a configuration file All information including the projection coeficiente will be stored in this file This file allows the stereo project to be reloaded later in time Project Name liet Tow Directory Je fdpryjet_flow chg Save Project The next steps to obtain stereo PIY data are Y Show NM anis File Name 1 Run the PIW MAP grid registration program 2 load a pair of stereo PIV images a valid projection is required lt Back Finish Cancel 58 PlVview User Manual StereoGeom Geometry Definition This stereo module is used to define the stereo viewing arrangement namely the camera positions with respect to the light sheet The entered coordinates are specified with respect to the same same reference point used by the P Vmap module Please note that in current implementation of P Vview the z 0 plane always defines the light sheet plane that is the X and Y axis lie in the light sheet while the Z axis points normal to this plane The camera positions can be entered directly in the fields on the right hand side The positions within the geometry plot are updated immediately and give the user a visual feedback Y ll Ps 2 z These tool buttons al
24. of the polynomial ratios is not always guaranteed Table 1 gives an overview of the associated number of coefficients and the required number of calibration points If the number of calibration points exceeds the minimum count a standard least squares fit is used to calculate the coefficients for the polynomial mapping functions a Levenberg Marquardt nonlinear least squares fit is used in the calculation of the polynomial ratios Table 1 Mapping functions and their Pe Mapping function Number of coefficients calloration RDA MAD Polynomial 1 Order Polynomial 2 Order Polynomial 3 Order Polynomial 4 Order Ratio 1 1 C e EEE E Ratio e Tsz Raio 27 T Ratio 27 2 Mapping from reconstructed image to object space Mapping between the reconstructed image and physical space is described by the following linear transformation Xin Xopj M X Offset where the magnification M is given in pixel mm y Obj y Offset Ym M PIVmap3 Input Menu The main menu shown Figure 5 consists of two input areas The left area is used to specify parameters that are common to all camera views and or all image planes The right area is specific to each supplied calibration image which is referred to here as plane In the Global Settings area the following parameters are specified PlVview User Manual Configuration data file stores a summary of all parameter settings within the current instance of the program
25. or speckle shift mode Menu bar The menu bar is used to perform any required operation within PIVview and contains seven sub menus File for loading images saving data setting preferences or exiting the program Image for basic image display related functions brightness look up tables zoom etc as well as image enhancement G PIVview User Manual PIV for performing PIV analysis defining PIV interrogation parameters or regions of interest Plot for configuration of plotting parameters or exporting a plot of the currently displayed data Windows to activate additional data display windows such as correlation plane 2 D and 1 D histograms or data probing menu Help to display P Vview version information or available program options Main window Once loaded PIV images are displayed in the main window Image pairs are displayed by by selecting the brightest pixels from both images The image pair can be interactively interrogated at any position using the left mouse button The interrogation parameters for cross correlation analysis are defined in various setup menus The results of the correlation analysis are displayed in the status bar at the bottom as well as in the form of a color coded vector defined in the Plot Vector setup menu Main window Lookup table A lookup table LUT is used to convert image intensities to screen intensities or colors The image above shows the currently selected LUT as a ve
26. several shapes during the creation of a polygon File Menu File Load image loads one of the supported PIV image types File New resets the image mask by removing all objects File Save mask image saves the currently displayed mask as a bitmap file BMP View Menu View Zoom in Zoom out changes the magnification factor of the image on the screen View Darken image Brighten image changes the brightness of the displayed image View Change LUT changes the image display lookup table View Toggle pixmap size displays larger tool buttons MEA 52 xr 00 PlVview User Manual 39 Draw Menu Draw Toggle current mask either calculates and displays the current mask by overlaying on the image or switches back to edit mode Draw Invert mask calculates the negative of the current mask Draw Move resize objects allows selection of objects and as well as their size modification Draw Move object vertices allows the shape of polygon objects to be changed Draw Draw rectangles creates a rectangle on the screen by dragging the mouse lts shape can be later modified using the resize objects entry Pressing the shift key allows the creation of squares Draw Draw ellipses creates ellipses Circles are generated by pressing the shift key Draw Draw centered ellipses creates ellipse starting from their center point Circles are generated using the shift key Draw Draw Polygons allow
27. strong stationary background artifacts which may hide the actual signal Correlation peaks falling within the specified radius from the origin are disabled 20 PIVview User Manual PIV Evaluation Setup Outliers Outlier Detection Methods The definition of outliers in experimental data can be very subjective In general an outlier in PIV data can be considered to be a spurious vector within an otherwise homogenous displacement field see Sec 6 1 Data Validation in Raffel et al 1998 for more information When activated several methods for outlier detection are available By combining the various methods most outliers are detected with very little valid data disabled Known also as optimization of cost Eu PIY Evaluation Setup Grid Correlation Algorithm Feak Detection Walidatior Conversion Filters etc Y Outlier Detection Outlier Replacement IZ Tw lower order peaks IZ Interpolation J T Max Displ Difference px 5 0 F Z Re evaluate with larger sample M Normalized Median Filter thresh 13 0 Imulti grid only 4 T Dynamic Mean Filter Mean 1 0 War 1 0 T Minimum signal to noise ratio 5 0 Ratio Peaki Peak 2 2 0 F Minimum correlation 20 f Global Histogram Filter Masimum number of validation passes 5 Apply Cancel Maximum Displacement Any displacement vector exceeding the specified magnitude is considered to be an outlier
28. transformation on the P Vview output using other software When this menu is activated a ruler and current origin are displayed within the image The ruler may be used to mark off Know lengths in the image The magnification factor can be interactively set by supplying the associated physical length in the menu Movement of the origin changes the entries in the menu accordingly gg PIY Evaluation Setup ES Interactive ruler Time Length Units Pulse Delay B US a 336 0 px Magnification factor JE Magn in ps mmn hsss3 A ee nm 2 Reset Ruler Displacement Unit Jum Velocity Unit ms o Origin Image Origin 378 y 322 pixel Origin Offset Az o ys o E o mm SB Process gt Defaults Apply Cancel OF Pulse Delay specifies the time delay between the two exposures in the image pair Generally does not apply to speckle displacment data Magnification Factor The scale factor used to conversion the image displacement to a physical displacement It can also be set interactively using the calibration menu Length Unit The units used for the physical coordinate system It is also associated with the magnification factor Displacement Unit The unit associated with speckle displacement data Also the physical particle image displacement prior to its conversion into velocity Velocity Unit The unit associated with the velocity as estimated by PIV
29. ttt ttttttt FERRE AAA AAA A dee e e e id 1 T t tt tt t JEND 1 E es L E a HHH ve eno ran ara op op ne 4 t r tttttrtrttr H H H BEER ttt ttt tt t 4 RERAR ttt t 4 4 Cee UIE An anlegt HHH Figure 4 Images of various calibration targets In all cases the camera is located to the right of the reference marks The program PIVmap3 This program is intended to provide accurate image object calibration of oblique distorted camera views as well as estimate the position of the camera if a series of Z translated images is available The core of the program is an algorithm to reconstruct a regular grid from images of a calibration target The following gives a brief overview of the reconstruction algorithm before the programs features are described in detail Grid reconstruction in PIVmap The reconstruction of the grid itself is a three step process Step 1 Image enhancement A variety of filters can be used to increase contrast of the image if the following two steps fail to produce reliable results Step 2 Feature extraction This is based on correlation based pattern matching that is a grid of dots is correlated with a template containing a single dot Similarly a template containing a cross is used to detect a grid of crosses or line grid As the correlation locks into features brighter than the 46 PlVview User Manual background images may need to be inverted a priori This is done automatic
30. 000 grid crossings Misalignment of the grid with respect to the light sheet may increase the measurement error Reference marks placed on the grid allow doubiless identification of image orientation and also help in obtaining an image object space referencing Important Make sure to check the invert option in case your image has dark lines on a bright background Otherwise PIVmapTO1 will not be able to properly detect the grid lines a EEE a E HE 1 e OR A fos r Fj 60 PIVview User Manual PIVmapT01 Toolbars amp Menu Options File Load image opens a file dialog for loading new calibration images File Load project opens and loads an existing project File Save project save current project Tools Place origin allows rough placement of the origin using the cursor The grid detection set the origin to the line crossing closest to this point Tools Complete evaluation performs full reconstruction on loaded image s using current parameter settings Tools Line detection performs line detection on current image using the current Hough transform processing parameters Tools Form mesh tries to reconstruct the mesh from the detected set of lines Tools Reconstruct grid detects the line crossing in the mesh and determines the major axis Tools Find optimal map provides an initial guess for the back projected image by trying to keep the magnification factors near unity Fine tuning is possible
31. 268 Wernet MP 2005 Symmetric phase only filtering a new paradigm for DPIV data processing Meas Sci Technol 15 2001 2028 Westerweel J Scarano F 2005 Universal outlier detection for PIV data Exp Fluids Online First publication Westerweel J Dabiri D Gharib M 1997 The effect of a discrete window offset on the accuracy of cross correlation analysis of digital particle image velocimetry Exp Fluids 23 20 28 Wieneke B 2005 Stereo PIV using self calibration on particle images Exp Fluids 39 267 280 Willert C 1997 Stereoscopic digital particle image velocimetry for applications in wind tunnel flows Meas Sci Technol 8 1465 1479 Willert C amp Gharib M 1991 Digital particle image velocimetry Exp Fluids 10 181 193
32. ABOVE EXCLUSION AND LIMITATIONS MAY NOT APPLY TO YOU Limitation of Liability TO THE EXTENT NOT PROHIBITED BY LAW IN NO EVENT SHALL PIVTEC BE LIABLE FOR PERSONAL INJURY OR ANY INCIDENTAL SPECIAL INDIRECT OR CONSEQUENTIAL DAMAGES WHATSOEVER INCLUDING WITHOUT LIMITATION DAMAGES FOR LOSS OF PROFITS LOSS OF DATA BUSINESS INTERRUPTION OR ANY OTHER COMMERCIAL DAMAGES OR LOSSES ARISING OUT OF OR RELATED TO YOUR USE OR INABILITY TO USE THE PIVTEC SOFTWARE HOWEVER CAUSED REGARDLESS OF THE THEORY OF LIABILITY CONTRACT TORT OR OTHERWISE AND EVEN IF PIVTEC HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES SOME JURISDICTIONS DO NOT ALLOW THE LIMITATION OF LIABILITY FOR PERSONAL INJURY OR OF INCIDENTAL OR CONSEQUENTIAL DAMAGES SO THIS LIMITATION MAY NOT APPLY TO YOU In no event shall PIVTEC s total liability to you for all damages other than as may be required by applicable law in cases involving personal injury exceed the original purchase cost of the PIVTEC software The foregoing limitations will apply even if the above stated remedy fails of its essential purpose PIVTEC RESERVES THE RIGHT TO MAKE CHANGES FROM TIME TO TIME IN THE CONTENTS OF PIVTEC SOFTWARE WITHOUT OBLIGATION OF PIVTEC TO NOTIFY YOU AS A LICENSE HOLDER OF SUCH REVISIONS OR CHANGES Acknowledgements The PIVTEC Software was and continues to be developed in close cooperation with the Institute of Aerodynamics and Flow Technology of the German Aerospace Center DLR in Go
33. HY TEC TEC PIVVIEW2C 3C Version 2 4 User Manual d Copyright 2001 2006 C Willert PIVTEC GmbH Sl G tti l PIVview User Manual PiVTEC GMBH Stauffenbergring 21 D 37075 G ttingen Germany Tel 49 0 551 205 29 170 Fax 49 0 551 205 29 176 49 0 2232 152849 Email info pivtec com Homepage http www pivtec com Version 2 4 15 April 2006 PIVview User Manual III Copyright 2001 2006 PIVTEC GmbH All Rights reserved worldwide No part of PIVview software or documentation may be reproduced transmitted transcribed disassembled reverse engineered stored in a retrieval system or translated in any form without the express written permission of PIVTEC GmbH Warranty PIVTEC GmbH PivTec warrants that the PIVview computer program and documentation will substantially conform to published specifications PIVTEC also warrants that the media CD ROM used to transfer the software is free from defects in material and workmanship and that the software is free from substantial programming errors for a period of six 6 months from the date of purchase unless a longer period is required by local law During this period defective media will be replaced and substantial programming errors in the software will be corrected by PIVTEC with no charge If PIVTEC is unable to replace defective media or correct substantial programming errors within sixty 60 days after notification of the defect or error PIVTEC will re
34. M I IMB f CAL T PPM e NC F DAT CDF Y Show Splash Screen Recent Image List determines how many image files should be listed in the File gt Recent Images menu option Recent File List determines how many PIV data files should be listed in the File gt Recent Files menu option Show Splash Screen if enabled shows the startup splash screen File Extensions This menu associates a variety of image types with certain file extensions These extensions are then used as file filters in various image selection dialogs The drop down selection offer the following image types PIV Images Pairs are used for cross correlation PIV and defines the first image of a pair of single exposed PIV images Double Exposed PIV Images are used for the selection of images for autocorrelation PIV Speckle Reference Images defines the zero state e g un deformed image used in speckle shift measurements Speckle Signal Images associates images where the speckles have been displaced with respect to the zero state image Mask Images are images that are overlaid and disable any data covered by them 32 PIVview User Manual Correlation Plane Window The correlation plane data is color coded using the currently active lookup table in the example above a gray scale LUT Bright pixels indicate a high correlation value dark ones a low correlation The local correlation values are displayed in the st
35. PlVview User Manual 33 Global Histogram Window This plot gives a visual representation of all displacement values available for the evaluated image Each colored point in the two dimensional plane represents a displacement vector whose coordinates correspond to the actual shift in pixels The boundary of this plot is defined by the size of the correlation plane with zero shift at the center The rectangular and elliptical regions near the center are user sizeable regions of interest These ROls can be used as a form of data validation any points on their exterior are considered to be outliers To use this function the Global Histogram option must be enabled in the validation menu Correlation Plane Yiew Qm Global histogram of PIV data with two ROIs Detail of the figure to the left Points enabled one rectangular currently outside of the ROIs are labeled as outliers selected and one elliptical The one as indicated by their red color dimensional PDFs are shown for the X axis below and for the Y axis on the right One Dimensional Histograms immediately bordering the 2 D histogram are vertical right and horizontal below one dimensional histogram of the two respective displacement components They are projections of the two dimensional histogram onto either of the major axes Their range therefore corresponds to the respective dimension of the correlation plane Sub Pixel Histograms these histograms locat
36. Spacing 2 mm Later the actual magnification factor Hari verica 2 Units mm is calculated from this Image splitting is required for images of the B16 file format which may contain a pair of images Further proper calibration image processing requires bright white Also some information regarding the image grid lines on a dark background The inverse black lines on white need to be type needs to be given For proper grid ee detection it is important for the program to know whether the image contains dark lines on a bright background or bright lines on a dark background Image Information B16 image treatment Image contains Image A No image split Black Lines 9 Preview A Image E No image split Black Lines r Preview B Step 3 Calibration Images Orientation Create new Stereo PI Project i Calibration Images Orientation In case the image needs to be flipped in A In order to properly match up the grids the images must have the same oreintation some way m Irror view this can be Below you can flip either of the images about the horizontal and or vertical anis achieved here IT Flip T lt gt B F FlipLe 2R Flip T lt gt B Show full scale Show full scale PIVview User Manual Step 4 Set Reference Point The buttons Display A and Display B show the currently loaded grid image Using the mouse the position of the origin can be selected The grid detection alg
37. The histograms on the below were obtained from the same image but processed with a different peak fitting algorithm least squares fit No peak locking can be observed the sub pixel histogram has a near uniform distribution Peak locking generally arises through a combination of certain particle image sizes typically too small whose narrow peaks in the correlation plane are analysed with the wrong type of peak fitting algorithm Another cause may be the non uniform spatial pixel response of the recording device CCD in combination with particle images with diameters less than 3 pixels One dimensional PDFs of PIV data Same PIV data as on the left but containing peak locking effects processed with a different peak fitting algorithm no peak locking is noticable PlVview User Manual 35 Batch PIV Processing a Batch processing is a means by which large number of PIV recordings can be evaluated in an efficient manner A common set of analysis parameters is used for all images During the processing the program accumulates the mean and RMS displacement and velocity data which are stored in a separate file after completion of the processing sequence a Batch PIY Processing E Directory ls piveJet3C Cam File Type Sinale expozed FI Images b16 ma bmp tit Filter 1 l r Filter 2 i Available Files Selected Files min irna b1 b pivC1_0001 b16 pivC1_0002 b16 pivC1_0003 b16 pivC1_0004 b16 eC DUR bE
38. The load function can retrieve previous parameter settings The parameter files are ASCII formatted and may be edited with any text editor 49 PIYmap3 Iof x Dir m Load n3C Stereo PlyYmap3D emo PipeFlow 3planes map S Target Views Type of target Number of views 2 Cameras Z translated images E Planes al Grid Spacing on Target x mm 2 Y mm 2 IZ Common grid for all views Type of target specifies whether the images of the calibration target contain dots crosses and lines Further the image characteristics with respect to features black on white or white on black are specified here This may be specific for each individual view Output Image Size Width px 600 Height px 500 a Figure 5 Main input menu of program P Vmap Number of views allows up to 4 camera views Stereoscopic PIV would require only 2 views A triple camera DGV setup would require 3 views A single camera view could be used to calibrate a standard 2 C PIV experiment Z translated images determines the number of calibration images for each view At least two are required if the camera position is to be determined Grid Spacing on target determines the physical distance of the calibration points on the target Output Image Size specifies the image size of the reconstructed de warped image This parameter is common to all camera views Calibration input area Loa
39. The tab dialog is accessed with the button and is split into an upper portion giving feedback on the current status of PlVmapTO1 such as which image is available and displayed or whether grid reconstruction has been successful The lower tab dialog described below follows the main processing steps of P VmapT07 The buttons A and B toggle the display between images A and B The parameters in the first two tabs are unique to each image a change accordingly The Projection tab parameters are identical for both images The Go Button starts a complete processing of the loaded images using the current parameter settings If successful the mapped image pair is displayed in the main window 62 PlVview User Manual PIVmapTO1 Dialog Detect Lines This tab is devoted to the line detection portion of T01 T01 C DPIV stereotest GridA tif 7 21xi based on the Hough transform The following c eee ON ee parameters can be set h Y Y e Pixel threshold threshold for pixels to be used in 44494 the Hough transform Proj Go High value less lines but faster Detect Lines Mesh and Map Projection Low value more lines slower ars e Negative Line threshold threshold for detection of the Pixel threshold een maximum in the Hough plane Line threshold Bere High value less lines Filter size low value more lines Scan size theta Scan size radius LHH LT d PL z dik dik dtk Filter size
40. V image has been evaluated It gives detailed information about any calculated vector within the data set The vector closest feo PlVview User Manual 9 to the mouse cursor is selected Statistics provides information about the currently displayed PIV data set such as mean and RMS values of displacement velocity vorticity etc Live Evaluation if enabled PlVview will display newly computed vectors during processing of the image Clear PIV Data deletes the currently displayed PIV data set Usually any modification in the PIV Configuration menu has the same effect The Plot menu Plot Settings opens a dialog in which the layout and properties of the plotted PIV data can be controlled Among these are vector coding and length scalar image display axis and axis labeling reference vectors and color definitions Shortcut key is P Subtract Mean displays the vector field with or without the mean displacement of the data set subtracted This is especially useful for displaying flow structures that might be hidden in a large nearly uniform convective flow Print Preview Printer Setup Print handles printing within PlVview and includes previewing layout and printer selection Shortcut key for printing is P Export as image stores the currently displayed data set or image in an image file screen capture Export formats include BMP JPEG PNG and others Current zoom factors are used in the exported i
41. age will be overlaid with a pair of coordinate axis herein referred to as Helper Axis These may be adjusted with the mouse and serve to define the principle directions in the grid Details on this are given in the following section The calibration area is accompanied by a tab menu containing the tabs Filters Scan Scale and Map Their purpose is described in the following Pl map3 Calibration View A File Mapping View Tools Help Io Su CAL Viena Pine 03 Eg GRID Qro IQ ve Filters Scan Scale Map P Medan ier 3a gt T Histogram Equalization T HighPass Filter Kernel width E T Binarize Threshold E q F LowPass Filter x Kernel width fos Apply Image 3121131 Value 737 Filtered 285 View Z 0 5mm c12 05 Ui Figure 6 Calibration view of program PIVmap3 with Filters Menu shown Filters Tab To improve the feature detection several filters may applied to the image They are shown top down in the order of their application Median Filter applies median filtering on the image Either a 3x3 or 5x5 kernel may be used Histogram Equalization performs histogram equalization on the image High Pass Filter applies a Gaussian weighted high pass filter to the image The width of the Gaussian filter function can be specified Values between 3 and 10 are recommended The processing speed decreases with increased width of the weight function
42. al in complex geometries involving windows or change of media i e air water interface Soloff et al 1997 introduced a mathematical formalism that combines back projection and reconstruction in a single step The procedure requires multiple images of the calibration target as it is translated at known displacements along the Z axis typical 5 7 recordings The drawback of this approach is that the camera position is not retrieved directly and therefore unsuited for DGV In effect the resulting projection equations connect object volume coordinates with planar image coordinates thus simplifying the vector reconstruction in stereo PIV A third approach to camera calibration originates from the field of photogrammetry and image vision Common here is the use of a pin hole camera model with added parameters to account for radial distortions as suggested by Tsai 1986 The calibration procedure involves a so called bundle adjustment of object space calibrated feature points using a least squares method In practice this can be achieved by translation of the target as in the previous case More general approaches can even calibrate on sequences of randomly oriented target recordings The major drawback of this method is that the camera parameters so called internal parameters such as pixel pitch need to be specified Specification of these intrinsic parameters is especially difficult when imaging fiber bundles are used Also this model does not
43. ally by specifying the type of target in the input menu e g black dots on white background In any case the image shown in the calibration menu should contain bright features dots crosses or lines on a dark background See sections below for further details Step 3 Grid reconstruction This step attempts to assign physical coordinates to the detected feature points The grid Origin is assigned to the point closest to the origin of the manually placed helper coordinate axis red and yellow arrows The helper axes are then used to find the 4 neighboring points on the major axes as well as 4 neighboring diagonal grid points The resulting group of 3x3 points forms the basis for the reconstruction of the remainder of the grid using a predictor corrector approach To allow for deviations of the grid nodes from their predicted positions a reconstruction tolerance is defined given in percent The reconstructed grid will have grid index coordinates assigned to the valid correspondence points Points on the major axes are shown in purple while unmatched points are shown in orange Reconstructed grid showing labeled grid points Reconstructed grid with X axis pointing down Unmatched points are shown in orange while Y axis is pointing to left General strategy on feature extraction The image sample size and the sampling overlap determine the detection efficiency of the grid As a rule of thumb the sample size should be near the node d
44. ated correlations PlVview User Manual 15 PIV Evaluation Setup Algorithm Advanced Correlation Techniques If none of the advanced algorithms are activated PVview uses the classical cross correlation method as described in various PIV references e g Willert amp Gharib 1991 More advanced algorithms are briefly described in the following Eu PIY Evaluation Setup Grid Correlation Algorithm Peak Detection Validation Conversion Filters etc Interrogation method Multigrid interrogation Grid refinement r Hulitple pass options Humber of iterations 5 2 Multi grid options Initial sampling window bi x b4 r Final window size 32832 px Image interpolation options Sub pisel image shifting Enabled on all passes r Scheme B Spline interpolation best r Order E Advanced options No speed optimization r Apply Cancel DK Interrogation method Allows the selection of different PIV interrogation schemes Standard single pass interrogation does classical PV interrogation without additional passes to improve the result Multiple pass interrogation The interrogation of the image is repeated at least once more In the following passes the image sample positions are offset by the integer shift determined in the preceding pass Once the residual shifts are less than one pixel a re evaluation of the respective point is no longer necessary e g convergence The maximum number o
45. ation mode Shortcut key is F3 PIV ROI puts PlVview into the ROI mode in which a rectangular portion within the image can be defined for processing To use this function the Use ROI option in the PIV Setup dialog has to be enabled Shortcut key is F2 Process Log shows a window with processing messages PIV Configuration opens a tabbed dialog in which a wide variety of PIV processing parameters can be defined Shortcut key is F4 or via context menu by right clicking the mouse Evaluate image performs an evaluation of the loaded image pair For large data sets a progress bar is displayed during processing allowing the user to abort the interrogation Shortcut key is F5 Batch Processing opens a dialog through which ensemble PIV processing can be initiated Correlation Plane displays the correlation plane and associated PIV image samples if chosen in a separate window The data in this window is updated each time a new interactive interrogation is performed on the image by clicking the left button within the image Shortcut key is F6 2 D Histogram PDF also is only available for evaluated PIV images It plots the calculated displacement data in a two dimensional plane and can be used in the vector validation procedure Additionally one dimensional histograms PDFs show the projection of the 2 D data onto each major axis Shortcut key is F7 PIV Data Probe is a window that is available after a PI
46. atus line by moving the mouse across the image Red lines at the center of the BS correlation plane marks the origin that G er R uf as d amp is the position of zero displacement gt Green Rectangle near the center of the correlation plane indicates the region of interest for the peak detection algorithm It is only visible if the ROI peak search is enabled in the processing setup Image Samples to the right of the correlation plane the two image samples from which the correlation was calculated are displayed if the toolbar button is enabled Note In case of double correlation the two images represent the two correlation planes from which the plane shown on the left was calculated Bo com 0 739 Change Image Magnification these buttons increase or decrease the magnification of the displayed image data Change LUT alters the lookup table LUT used to display the correlation image data Each time the button is pressed a new LUT is applied Display of Image Samples this button must be enabled to display the image samples in addition to the correlation plane Saving the Correlation Data the currently displayed correlation data and image samples can be stored in ASCII files by pressing this button Once the data is saved the button is deactivated until new correlation data becomes available Printing of the correlation plane using printer selected in Plot Print Setup menu option
47. available some of Kodak ES1 0 K l I010 r which reguire additional input regarding the type of camera sensor with which the images were acguired JA Estimate Map Tab Parameters specified in this menu determine image mapping coefficients that perform the back projection from camera view to an un distorted image The choice of mapping function is specified on the top of this menu Magnification PIYmap Control Es determines the final magnification factor for the tee een Ges he reconstructed image This value is constant throughout Projection Map Order 2 y the Image ey ET Translate image Translate image By default the origin of the grid will be shifted to the lower SN mm 21 Z left corner These parameters allow the origin to be Shitty mm 4 shifted LD Copy to all views Copy to all views i x Fit to output image copies the current parameters to the remaining camera views This is especially of interest when all reconstructed views should coincide after mapping them Fit to output image is useful to approximately match the reconstructed grid to the specified output image size and should be used on a new set of calibration images An even better match is achieved by manually fine tuning the magnification and shift parameters Dewarp this view calculates the mapping coefficients for each reconstructed grid of this camera view These mappings are used to back project the original images which are then di
48. ble line segments are displayed PlVview User Manual PIVmapTO1 Dialog Mesh Map 63 This dialog contains all parameters necessary for reconstruction of the grid from the previously detected line segments Form mesh button connects line segments to form continuous grid lines The result is displayed by overlaying the image with green lines Mapping button calculates the back projection mapping coefficents Scan size image search distance between neighboring line segments Default of 5 usually does not need to be altered Theta tolerance maximum possible angle degrees allowed between two neighboring line segments For strongly curved grid lines this value may need to be increased default 5 Minimum count minimum number of line segments necessary to form a grid line high value less short lines less chance of outlier lines low value permits more short lines In case of outlier lines just increase this value Base angle defines the principle direction of the X axis An angle of 0 corresponds to having the X axis run horizontal from left to right In case of diagonal lines choose a value of either 45 or 45 TOL C DPT stereotest GridA tif IE xl Image Loaded Lines Mesh Map Ivvv S WW VW Poi Go Detect Lines Mesh and Map Projection Form mesh Mapping Mesh forming Scan size image 5 Theta tolerance 15 Minimum count 5
49. coordinates x y the first order projection equations are _ AX Ag Y Hua za aog AxyX 7031 Y A322 A33 As TO oT inks Taig Az0X T Az Y T A322 T Az with a 1 Although these equations do not include higher order distortions they nonetheless are sufficient for most PIV applications and show deviations on the order of 1 10 of degree with respect to the more precise pinhole camera model by Tsai 1986 These higher order methods are also included in P Vmapg Light sheet plane b Light sheet plane object plane object plane at z 0 at z 0 image plane CCD sensor i image plane CCD sensor Figure 1 Coordinate system and viewing arrangements Left co planar viewing arrangement with object plane parallel to image plane Right non co planar off axis arrangement with image plane tilted with respect to object plane a Grid imaged without b Grid imaged obliquely in a c Grid imaged with strong lens distortions on co planar non co planar arrangement distortion arrangement Figure 2 Examples of imaged grid under different viewing arrangements PIVview User Manual 45 A Reference plane at z 0 am Figure 3 Multiple plane calibration procedure based on translation of calibration target Camera A and B comprise a single sided stereo configuration Camera A and C view the light sheet from two sides Here the light sheet thickness should be accounted for t tt
50. ctical Guide Springer Verlag Berlin ISBN 3 540 63683 8 Astarita T Cardone G 2005 Analysis of interpolation schemes for image deformation methods in PIV Exp Fluids 38 233 243 Hart D P 2000 PIV error correction Exp Fluids 29 13 22 Brent RP 1973 Algorithms for minimization without derivatives Prentice Hall Englewood Cliffs NJ Ch 5 Coudert SUM Schon J P 2001 Back projection algorithm with misalignment corrections for 2D3C stereoscopic PIV Meas Sci Technol 12 1371 81 Meinhart C Wereley S Saniago J 2000 A PIV algorithm for estimating time averaged velocity fields J Fluids Eng T ASME 122 2 285 289 P renne N Foucaut JM Savatier 2004 Study of the accuracy of different stereoscopic reconstruction algorithms In Stanislas M Westerweel J Kompenhans J eds Particle image velocimetry Recent improvements Proceeding of the EUROPIV 2 workshop Zaragoza Spain March April 2003 Springer Berlin Heidelberg NewYork pp 375 389 Press WH Flannery BP Teukolsky SA Vetterling WT 1988 Numerical recipes in C the art of scientific computing Cambridge University Press Cambridge Ch 10 2 pp 299 Tsai RY 1986 An efficient and accurate camera calibration technique for 3D machine vision Proc IEEE Conference on Computer Vision and Pattern Recognition CVPR 86 Miami FL pp 364 374 76 PIVview User Manual Wereley ST Meinhart CD 2001 Second order accurate particle image velocimetry Exp Fluids 31 258
51. d Images permits the simultaneous selection of all calibration images associated with this camera view The common directory of these files is specified above the table Table Area The table itself lists the image files their position in Z and the specified reference plane Clicking on the image file shows a preview of the corresponding image in the area below the table The out of plane translation Z position of each image can be entered directly in the corresponding cell The reference plane indicates that the camera position and image mapping is calculated with respect to the given image and is set by clicking on the corresponding cell Single images can be loaded by clicking the icon on the far right of the table Mask image can be used to mask out areas in the calibration images to improve grid reconstruction performance and also increase overall calculation speed less correlations required Mask inversion disables the complementary area A common mask is used for all images of a single view 50 PlVview User Manual PIVmap3 Calibration Area Once a sequence of images has been loaded they may be displayed one by one in the calibration menu The display should display the image with bright features dots crosses on a dark background If not then the image type was not properly set in the main menu i e black on white vs white on black Also any optional filtering that has been applied to the image will be made visible here The im
52. d kernel of specified width Differentiation Calculate cell size for differentiation The estimation of differential quantities e g vorticity shear strain etc relies on calculating differences between neighboring vectors and can therefore be very noisy especially if the seeding density is low and the data points are closely spaced Generally it was found that differentials calculated from PIV data sampled at less than 50 overlap are increasingly noisy due to artifacts introduced as a consequence of nearly identical particle image pairs in neighboring sampling positions In other words for correlation planes calculated from two nearly overlapping sampling areas most of the particle image pairs can be found in both samples Especially in areas of high gradients the data will be biased according to the mean displacement of all pairs Closely sampled data will then show a similar bias in neighboring points and thereby provide an underestimated differential When activated this option checks the sampling interrogation window overlap and from it determines an optimal distance from which to calculate the differentials Alternatively the user can specify the spacing for the calculation of the differentials Extrapolate to edges Generally differentials cannot be estimated at the borders of the data set However by extrapolating the vectors beyond the edge using the interior vectors the differential at the edges can be estimated 24 PIVvi
53. d width can be specified here Also there is the choice to fill in the arrow heads Vector Stretch Factor Specifies the amount the displacement is scaled with respect to the image For example a vector with length 10 pixels with a scale factor of 2 0 is 20 pixels long when plotted on the image The stretch factor is always based on the pixel displacement of the data Maximum Vector Can be used to avoid plotting vector whose length exceeds the value specified e g outliers The magnitude is specified in pixels Spacing This option allows vectors to be skipped and is useful when the data is too dense to be properly displayed Outlier display Determines how outliers should be treated on the screen Mark outliers with X when activated any vector labeled as an outlier will only be indicated by an x Hide Outliers when activated any vector labeled as an outlier will not be shown Show Outlier Vectors when activated any vector labeled as an outlier will be shown Offset Vector Subtract Image Offset When activated the image offset specified in the evaluation setup menu will be subtracted from each displacement vector Subtract Mean Displacement When activated the mean displacement of all vectors in the field will be subtracted If disabled a user specified offset can be subtracted see below 26 PIVview User Manual Specified Offset Vector The user can specify a given displacement to be subtracted fr
54. e active if image or data is available PIVview User Manual 7 The PIVview Menu Entries The File Menu Evaluation Mode specifies one of the three P Vview operation modes gt Cross Correlation Mode the standard mode for PIV analysis based on double frame single exposure images image pairs gt Autocorrelation Mode PIV analysis based on single frame double exposure images which are common in photographic PIV gt Speckle Displacement Mode for the analysis of speckle displacement image pairs This mode is similar to the cross correlation mode but instead of converting the data to velocity it is presented as displacements Also the two required images are loaded separately Preferences opens a dialog which allows to set a variety of P Vview global settings such as file extensions default data output etc Open image opens a file dialog for loading a PIV image pair or speckle shift image Open data opens a file dialog for loading a PIV or speckle shift data set Save data set opens a menu with output options to save computed PIV data in a file The Save menu option uses the file title without extension to create a name for the output file and places it in the same directory with the original image With the Save as option the user can define the name and destination directory of the output file Create Tecplot Data to save the processed data set in a format compatible with Tecplot Style files for Tecplot ar
55. e also generated Parameters Load Save Save as stores or retrieves the current P Vview parameter configuration in from a file This file is formatted in netCDF and therefore is platform independent but cannot be read be a standard text editor Recent images Recent files are listed by their names and are opened by clicking on them Note that the PIV mode of operation does not automatically change for different image types New Window starts are another instance of P Vview by opening additional main window New windows operate completely independent of one another Exit terminates P Vview program The current PIV image and plotting parameters are stored in the Windows registry and are restored when the program is started again Tip Use parameter files if you plan to work with different types of images The mage Menu Be Image PreProcess opens a dialog in which a variety of image pre processing options can be set such as splitting of the PCO camera s double image files image filtering histogram adjustment background subtraction etc Eat Edit Mask starts the mask editor P Vmask with the currently loaded image Completed t mask images can be loaded directly from in the image pre proceesing menu Calc Minimum Image allows the selection of an image sequence from which a minimum intensity image is computed Horizontal Profile Vertical Profile shows plots of image intensity profiles for the current
56. e located under HKEY_LOCAL_MACHINE Software PivTec PIVview2 Removing PlVview The best way to remove P Vview is to use the Uninstaller which is part of the Windows operating system found under Configuration Software 4 PIVview User Manual Alternatively the software is completely removed by deleting the P Vview installation directory This however does not remove the registry entries associated with P Vview PlVview User Manual 5 Getting acquainted with P Vview The user interface of PlVview is similar to many common Microsoft Windows based applications and features atitle bar amenu bar several toolbars a main window to view data workspace a Status bar PlVview is based on a single document software architecure that is only one image pair can be opened by each PlVview instance If you need to open several PIV images simultaneously you need to start several instances of P Vview which can be started through the menu option File New Window Please note that the settings of last P Vview instance closed are stored in the registry Mask Edit ma a ris Main window of PivView with image pair loaded and interactively sampled interrogation vectors The intensity LUT is displayed on the left The top left shows a reference scale vector Title bar The title bar displays the file name s of currently loaded image or data file s and PIVview s current mode of operation e g cross correlation autocorrelation
57. e window for interactive image probing opens the configuration menu shown to the right of image view performs a full interrogation of the image pair followed by data validation The result is displayed as a green vector map applies a least squares fit to the validated data and displays a red vector map of the fit saves the coefficients of the current least squares fit and prompts the user whether to enable the fit in which case P Vview reloads the image set with modified projection coefficients Once the alignment map has been activated this window can be used to verify the actual alignment Further alignment iterations are not possible In earlier versions of P Vview the misalignment correction between the camera views was referred to as residual alignment DD PlVview User Manual Some notes on performance In some instances the particle images viewed by one camera are only weakly distinguishable on the other This is frequently the case for air flows seeded with small particles utilizing a forward backward scattering arrangement For this reason rather larger interrogation windows are necessary When in doubt we suggest increasing the interrogation window until consistent displacement fields are produced The correlation plane window is also useful for this optimization The correlation module of alignment procedure actually makes use of all four images of a stereo PIV recording and uses a multiplication of correlation planes to en
58. ed the furthest from the 2 D histogram are generated by plotting only the non integral part of the displacement values Their range is from 0 5 to 0 5 with a default bin width of 1 20th a pixel Sub pixel histograms are especially useful to check whether peak locking effects are present in the data Zoom in Zoom out these buttons increase or decrease the magnification of the displayed histogram data This button enables the global histogram validation Alternatively the Outlier Configuration menu can be used toggles the currently selected ROI between elliptical and rectangular shape Hi BB 0 34 PIVview User Manual Validation with new ROI if global histogram validation is enabled then a change to the ROI can be tested directly on the data with this button Default ROI this function resets the currently selected ROI such that it nearly covers the entire histogram plane This is especially useful if the ROI is lost off the histogram plane Adds or removes ROls from the histogram plane Up to hree ROls can be activated 1 D PDFs and Peak Locking The histograms PDFs below illustrate the effect of peak locking which is a tendency of the peak fitting algorithm to lock into integer values The histogram on the top shows the presence of peak locking in the form of integer spaced peaks The corresponding sub pixel histogram below shows a triangular shape with increased counts near integer values middle of histogram
59. er description lt fileIN gt input netCDF data file as generated by PlVview lt file0UT gt output Tecplotformatted output file Command line options asc creates ASC II data file bin creates Tecplot binary data file vel writes out the velocity data contained in the input file if it is available This is the default output option disp writes out the pixel image displacement data contained in the input file piv writes out the pixel image displacement data contained in the input file mag also writes out either velocity magnitude or pixels shift magnitude depending on which option was set raw skip post processing of data after loading i e validation diff stores differentials implies ve1 dvort stores vorticity data implies vel dshear stores shear strain data implies vel dnstrain stores normal strain data implies vel PIVview User Manual 71 s lt num gt sets output stride to lt num gt that is only every lt num gt th point is stored sx lt num gt sets horizontal output stride to lt num gt sy lt num gt sets vertical output stride to lt num gt zZ stores Z coordinate z lt num gt stores user specified Z coordinate info provides feedback during processing of data NCDUMP The NCDUMP program reads the specified data file and writes the contents onto the command shell screen Calling procedure ncdump lt fileIN gt Parameter Description lt fileIN gt input any netCDF
60. ereo PIV Procedure in PIVVIGWSC u a ea ae 40 Vector reconstruction in stereo PIV eee 41 Enabling the Stereo Mode am ne ein 41 Stereo Menu Options amp Toolbar sinice e E 41 Creating a new stereo project With PIVmap3 u000000000nnnnnnn nenn nnnnn nenn nnnnnnnennnn nn 43 Background On Stereo PIV ua ee ee 43 INeprogran VIN e ae cast E on s a O S GEK 45 Ens eels enel Rie PIU ei cin 45 Calculation of the Mapping FUNCTION oocccncccoccnncononcnncononcnnnnnnnonnnnnnncnnnnnnnnnnonnancnnnonanens 47 FLV MAD O aE MENU E E E tn eier 48 PIVIMADS Vallbralion Aea 50 FEDETT AXIS se A ete Ha R s t t 53 Description of the P Vmap3 menu entries ee ee ee ee e e e ee 53 MENU MIC L 53 Men Mapping sra 54 MONU 9 TTT 54 Creating a new stereo project with StereoWizard css ssss sse ss cesse ssc sse seene eee 56 StereoGeom Geometry Definition seere ene a a i S E Eea EEA 58 PIVmapT01 Projection Module uzuu2002u20002an00nnnnnnonnannonnannunnnnnnnnnnnnnnnnnnunnnnnnnnnnn 59 Some notes on obtaining calibration grid Images sse eee eee eee eee 59 PIVmapTO1 Toolbars Menu OptionS sss see eee eee 60 P1VvMa9 TOT Ta Ee 2 at eaten ae 61 PIVMap TOT Dialog Detect RIL 62 FIVMap 01 Dialog Mesh 8 Map 63 PIVMap TOT Dialog PIOjeCllOn casona A S 64 Stereo View Disparity Correction sss sese ss sese ss sese ss sese ss sees seene ennenen nenen 65 Disparity Correction TOO Ste 65 Command Line Utility Prog
61. ers in a user specified project file This ASCII file is editable with any text editor i e Notepad The default extension is map File Exit leaves the program after asking the user for confirmation Menu View View Zoom out in changes the image display scale factor View Brighter increases the image brightness on the screen Short cut key B View Darker decreases the image brightness on the screen Short cut key D View Auto scale sets default intensity on image View Change color LUT skips to the next available lookup table LUT to convert the grayscale images to pseudo color 54 PIVview User Manual T Menu Mapping SA Mapping Input menu changes to the main input area of the program CAL Mapping Calibration mode a changes to the calibration area and displays the filtered image of the currently selected view and plane ren Mapping Scan image R performs the feature extraction scan by the localized template matching scheme using the current parameter settings Any manual change to previously detected feature points are lost u Mapping Create grid i tries to reconstruct the grid using the currently detected feature points ws Mapping Show mapped image a calculates the mapping coefficients for each reconstructed grid of this camera view These mappings are used to back project the original images which are then displayed The overall speed of the mapping process depends on the output
62. ettingen Germany Copyrights and Trademarks netCDF network Common Data Format Data Format Software Library and Utilities 1993 2005 University Corporation for Atmospheric Research Unidata All right reserved http www unidata ucar edu Tecplot Amtec Engineering Inc Microsoft Visual C Windows Windows Metafile Microsoft Corporation All other product names are trademarks or registered trademarks of their respective owners IV PIVview User Manual Contents OVErVIEW 9 PIVVIEW tontas 1 B CKIrOUN G T 1 IL 1 WIP COMMING TSATUTES use diras 1 What PIVAEWIS TNO aaa ee ae 1 PIUMIOWEN Sd a ee 2 Gredlls s ee di 2 System Requirements 2 2 na een 3 Teen ET e MEMO Ares rennen 3 TrsT Enie sse T 3 VWh tis installed Where T 3 Removing FIV VICW sia id 3 Getting acquainted With PIVVISeW ssss sss ss sss ss sese ss enesenn ennenen eenn nenen 5 TUG Dl TTT 5 MENUDO een erde 5 Malin es AAA A PP ee 6 Main Window Reference Vector eee eee eee 6 Main Windows Sro Bar S rearen ds 6 SAUS B Ii ee dl Ti 6 The PIV VIEW Men Entes aiii aa a aaa a Se aaRS 7 The Eile Mesina E ibi 7 Me Mage Men Usina ai cada 7 A A T 8 bH s OP NIG TTT T 9 NG WINdOWMen en ee 9 TG eel IE 9 The PIVVIeW Lele eT 10 The PLV VIEW Dialogs HT 11 The Image Pre Processing OLT 11 ni SSE SE t E SNEK 11 TaD PIllelSs nase E een Sees eignen 11 Tap Background seien 12 o unse eier 12 PIVAEVAlUallon Sellps un ee Eee 13 PIV Evaluation Setup
63. ew User Manual Region of Interest This display mode can only be activated if the ROl option is checked in the Grid Tab of the PIV parameter menu The region of interest ROI is used limit the evaluation grid to a defined rectangular region This region is shown as a red white border whose corners and mid points can be translated by holding down the left mouse button Yellow s inside the ROI indicate the centers of sampling positions for the correlation analysis Their spacing is defined in the Grid tab of the PIV configuration menu Image overlaid with ROI selection rectangle Yellow crosses mark the centers of the interrogation windows PlVview User Manual 25 Plot Configuration Menu ES Piv iew Plotting Parameters VWechors Contours Axis Ret vector Fonts Colors IZ Show vector Field Default Yector Size Spacing Stretch Factor a Max Vector px fio Spacing h 2 h 4 Outlier display Mark outliers with x gt Offset Vector No offset vector e A el mams Jo ms ma fo L Color Coding Validation has E Eolor Lookup Table Lu F Show LUT Vector style Vector Plotting Parameters Show Vector Field has to be enabled to show the vectors If a data set is available that is the image is evaluated then enabling this option will immediately display the vector field Vector Size Spacing Both the vector head length an
64. ewarping opens a menu allowing a series of images to be back projected using the PIVview User Manual 55 current image maps The image interpolation function and destination directory may be specified Tools Show log opens a window with past error messages warnings and comments 56 PlVview User Manual Creating a new stereo project with StereoWizard The menu entry Stereo New project calls up the program StereoWizard which allows a step by step definition of a stereo projection including the grid registration procedure Step 1 Calibration image Create new Stereo PIY Project Calibration Images Here the user is asked to either load a pair of calibration grid Images or an existing project A new stereo PI project requires a pair of he wishes to modify calibration images These should basically consist of grid lines of known spacing similar to the image on the right Proper processing requires white lines on a dark background Dark lines on a white background are also fine if specified as such Selection Calibration of Grid Images Pick File amp E OPIV StereoJet gidC1_0001 b16 Pick File B C DPIStereaJet gridC2_0001 bIE Import existing project Show Maps Help Cancel Step 2 Calib Image Basic Information Create new Stereo PI Project Calibration Images Basic information Using the dropdown entries give the actual physical spacing of the line in the grid here Grid
65. f PIVmap The images a overlaid with each other Only A calculates back projected image only for view A Only B calculates back projected image only for view B Origin ox oy x y coordinates pixel of the image origin within the back projected image Grid spacing x y pixel distance between grid lines in the back projected image This value should chosen to be close to the line spacing in the Original images Grid spacing x y mm distance between grid lines in physical space The ratio of this value and the pixel grid spacing gives the image magnification factor also shown in dialog Image Size x y pixel dimensions of the resulting image which is shown as a red box in the main window TO1 C DPIY stereotest Grid A tif E ix Image Loaded Lines Mesh Map Iv dd se vV vVVv Poi 648 G Go Detect Lines Mesh and Map Projection Calculate Only amp Only B Best guess Output Image Origin Grid spacing pr 40 2 y px 40 Grid spacing mm 5 physical u mm 5 Magn primm M v My 8 Image Size width 570 Height 480 4 Rotate deg pa DU Z 4 La dtk PlVview User Manual 65 Stereo View Disparity Correction In practice it generally is not possible to perfectly align the calibration target with the light sheet As a result the sampling windows of the PIV interrogation do not coinc
66. f interrogation passes can be specified Two to three passes are generally sufficient for convergence Multiple pass interrogation not only increases the data yield due to the higher amount of matched particle images but also reduces the so called bias error Westerweel et al 1997 Outlier detection is used to validate intermediate data sets Multigrid interrogation uses a pyramid approach by starting off with larger interrogation windows on a coarse grid and refining the windows and grid with each pass Outlier detection is used to filter the intermediate results Multiple pass options Number of iterations specifies the number of passes during multiple pass PIV interrogation Multi grid options Initial sampling window specifies the starting interrogation window for multi grid PIV interrogation The final window size is set in the Grid menu tab 16 PIVview User Manual Image interpolation options Sub pixel image shifting enables sub pixel based image shifting image deformation on multiple pass or multi grid interrogation This technique is a second order method and involves a complete deformation of the image data using the displacement data of the previous interrogation passes Interploation scheme provides a choice of different pixel interpolation schemes as well as the order In general cubic B splines were found to produce the best results Speed optimization options To increase processing speed intermediate p
67. fund the license fee to the licensee There are your sole remidies for any breach of warranty Disclaimer of Warranties YOU EXPRESSLY ACKNOWLEDGE AND AGREE THAT USE OF THE PIVTEC SOFTWARE IS AT YOUR SOLE RISK AND THAT THE ENTIRE RISK AS TO SATISFACTORY QUALITY PERFORMANCE ACCURACY AND EFFORT IS WITH YOU EXCEPT FOR THE LIMITED WARRANTY ON MEDIA SET FORTH ABOVE AND TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW THE PIVTEC SOFTWARE IS PROVIDED AS IS WITH ALL FAULTS AND WITHOUT WARRANTY OF ANY KIND AND PIVTEC HEREBY DISCLAIMS ALL WARRANTIES AND CONDITIONS WITH RESPECT TO THE PIVTEC SOFTWARE EITHER EXPRESS IMPLIED OR STATUTORY INCLUDING BUT NOT LIMITED TO THE IMPLIED WARRANTIES AND OR CONDITIONS OF MERCHANTABILITY OF SATISFACTORY QUALITY OF FITNESS FOR A PARTICULAR PURPOSE OF ACCURACY AND NON INFRINGEMENT OF THIRD PARTY RIGHTS PIVTEC DOES NOT WARRANT THAT THE FUNCTIONS CONTAINED IN THE PIVTEC SOFTWARE WILL MEET YOUR REQUIREMENTS THAT THE OPERATION OF THE PIVTEC SOFTWARE WILL BE UNINTERRUPTED OR ERROR FREE OR THAT DEFECTS IN THE PIVTEC SOFTWARE WILL BE CORRECTED NO ORAL OR WRITTEN INFORMATION OR ADVICE GIVEN BY PIVTEC OR A PIVTEC AUTHORIZED REPRESENTATIVE SHALL CREATE A WARRANTY SHOULD THE PIVTEC SOFTWARE PROVE DEFECTIVE YOU ASSUME THE ENTIRE COST OF ALL NECESSARY SERVICING REPAIR OR CORRECTION SOME JURISDICTIONS DO NOT ALLOW THE EXCLUSION OF IMPLIED WARRANTIES OR LIMITATIONS ON APPLICABLE STATUTORY RIGHTS OF A CONSUMER SO THE
68. h sample produces an estimate of a feature W Arrow tips mark neighboring points point location adjoining samples will detect the same LB a ede feature point more than once This value specifies the a E minimum number of times a feature point needs to be F Elliptic region of interest detected to qualify as such The merging radius given below has a direct influence eee aaa Merging radius Eisen specifies the scan area to combine near by feature ER Build Grid point estimates Values between 3 and 10 are recommended Reconstruction tolerance is required to reconstruct the grid from the detected feature points The reconstruction is based on a predictor corrector approach where valid grid points are used to estimate the position of neighboring grid points The search area is determined by this tolerance value Values between 5 and 10 percent are recommended Arrow tips mark neighboring points is used to aid the reconstruction algorithm in finding the first points near the origin Limit range of nodes restricts the range of nodes in the reconstructed grid This is useful if the image has many invalid feature points on the edges The use of a mask specified in the main menu should however be favored Circular or elliptic calibration geometries can be treated as well Edit move points manually In some cases the grid reconstruction may fail if falsely detected or missing feature points near the grid origin mislead the reco
69. hance the correlation peak The alignment procedure allows a choice on which of the two projection maps are modified the default is a symmetric adjustment of both views c PI iew 2 1 Inter camera alignment l 2 2 5 x Image Align Ar REAR a figye A e 3 Wabdation parameters Max Displ Diff 9 T Min Correlation T Peak to Peak Ratio AI eee ee 7 Display parameters afa gt m P K Ea BE Pos 128 160 dx 9 34 dy 0 11 Fit dx 9 96 dy 0 19 S Disparity correction window after least squares fit to offset data PIVview User Manual 67 Command Line Utility Programs The PilVview program suite includes several command line based utility programs that can be used for batch processing purposes debugging data retrieval etc These programs should be run from the command shell e g DOS box and their path PIVview Bin Or PIVview Utils should be added to the systems PATH variable These program are useful for batch processing large amounts of PIV images without using PIVview itself frequently done on Linux Unix systems AVERAGE This program first builds a list of files from the first and last file provided and then calculates Statistics from these The result is placed in a new file whose name is determined from the common file stub of the input files It is useful in the calculation of quantities such as mean velocity field or Reynolds stresses Calling procedure
70. hysical coordinates may be used The appearance of each axis may be controlled independently ES Piv iew Plotting Parameters Tl x Vectors Contours Ariz Ref Wector Fonts 7 Colors Y Show Ares Defaut Ans Labeling E Image Coordinates amp Physical Coordinates TANS e Show Labels pixel Skip Tic Spacing Tic Anchor 2 AIS e Show Labels pixel Skip Tic Spacing Tie Anchor Minor Tics Minor Tics T Show Title Title E TF Show Title Title lv Show Axes has to be enabled to place axes on the edge of the image or plot Image coordinates plots the data in image space pixels Physical coordinates plots the data in physical object space using the currently active scaling factors Show Labels has to be enabled to place labels on either of the axes Tic Spacing defines the increment of the tics to be labeled major ticks Tic Anchor sets a virtual starting point for the major ticks to ensure a nice layout with even label numbers Skip defines which of the major ticks should not be labeled 0 for to label all major ticks 1 to label every other etc Minor Ticks sets the number of minor ticks between two major ticks Show Title has to be enabled to place a title on either of the axes The title is defined in the entry field below 30 PIVview User Manual Reference Vector EE Pirv iew Plotting Parameters MEZ Vectors Contours Axis Hel
71. icient color codes the image according to the correlation coefficient from the displacement estimation Enstrophy color codes the squared magnitude of the out of plane vorticity o Lambda operator color coding according to Ou Ov i Ou Ov Ou v AS ez l Ox Oy Ox Oy Oy Ox Vector Count color codes the number of validated data in averaged PIV data sets only for averaged data 3 C vector residuals color codes the residuals from a 3 C vector reconstruction for stereo PIV only available for stereo PIV data other options may also be available Contour Display Range Specifies the range of contour values to be shown These values are also displayed on the lookup table above the contour plot Nice Range calculates nice or round limits for the color bar using the maxima available from the current data set Show Color Bar This option determines whether the reference lookup table is shown This color bar is placed above the data field Discrete Contours when enabled discrete contour levels are shown Otherwise a smooth transition from minimum to maximum is used Reverse LUT when enabled reverses LUT Wrap LUT starts from the lower end of the look up table for values exceeding the specified range Fill Background fills the image background with the zero contour color PlVview User Manual 29 Axis Plotting Parameters This dialog controls the type of axis and axis labeling on the displayed data Either image or p
72. ide between the two views One way to account for this is to perform a PIV correlation analysis on the particle images acquired at the same instant from the two views that is correlating an image from view A with an image from view B This correlation results in a vector map showing the misalignment between the two camera views This disparity misalignment map can then be used to adjust the projection coefficients PlVview3C offers this alignment procedure in the Stereo Menu entry Disparity Correction which opens the window shown on the right provided a stereo project and image set has been loaded The alignment window shows an overlayed image of the two camera views The mouse can be used to interactively correlation probe the image similar to the main P Vview window PIVview 2 1 Inter camera alignment Sa sf Bf 72 PI view 2 1 Inter camera alignment ml l Image Align Image Align i ara LE z Q u a ES S ae Evaluation parameters and Sample size Sample overlap Ea lt Overlap 32 px A Pos 297 7 Imgl 20 6446 Img2 59 5019 A Main window of the PlVview3C stereo Displacement vector map showing the camera disparity correction misalignment between the camera views Disparity Correction Tool Buttons The functions of the view tool buttons are equivalent to those of other views so only those unique to this window are described here shows the correlation plan
73. iew automatically determines the image name of the second image using the name of the first Several rules are used to determine the second image ima gt imb Ifthe first image has the extension ima then the program will look for a file where the extension changed to imb Example 1 image myImage ima gt 2 image myImage imb a gt b If the first image has the letter a before the extension then program will look for a file changed where a is replaced by b Example 1 image myImage_a tif gt 2 image myImage_b tif n gt n 1 If the images are sequentially numbered the program will look for the file with the next highest index Example 1 image img099 tif gt 2 image img100 tif Excepted from these rules are images loaded in the speckle or autocorrelation modes as well as images that are loaded as self contained image pairs e g PCO B16 double images PIVview User Manual 75 Supported Output File Formats ASCII Plain ASCII data can be generated by PlVview which can then be read by other programs without much effort A short header specifies the number of points in the data set along with a description of each column s content Tecplot PlVview data can be exported to be compatible with the TecPlot graphics package ANIT EC Engineering Inc www amtec com Version 7 0 and higher Style files for both displacement and velocity data are generated to ease the data presentation wi
74. image size and the number reconstructed planes m Mapping Process current view performs grid reconstruction and image back projection for all images in the currently selected camera view Mapping Overlay mapped images computes a combined image of all available back projected views for the currently selected image plane At least two valid back projected camera views are required This is especially useful to check the coincidence of the back projections 3 Mapping Process all images processes all images for all views including camera position estimation Mapping Center helper axis This utility re centers the helper axis to the middle of the image Mapping Copy helper axis This utility copies the helper axis from the previous plane to the current plane Mapping Estimate camera position tries to calculate the camera position using at least two calibrated planes of the current view A valid image mapping Is required The result even for non convergence is placed in the calibration tab menu Mapping Save calibration points exports the detected and validated feature points for all available plane of the current view to a user specified ASCII file Mapping Export for P Vview stores the current image projection coefficients along with all relevant data i e image size camera positions etc in a user specified P Vview3C project file If an existing file is selected the new data is merged into it Menu Tools va Tools Batch image d
75. istance in the image In general one should start with a larger sampling window If not all grid points were found change to next smaller sample or increase sample overlap If grid points were detected more than once increase the merging radius to combine neighboring points Depending on the tolerance the grid reconstruction procedure may ignore these misaligned calibration points Effect of sample size on dot detection The following images illustrate the effect on sample size on the feature extraction Detected dots are labeled with a cross PlIWview User Manual 47 0 eo Pra e e0eee zu 9 46 0 o 4 40000 TEO e rend HH 8 4 4 HH 4 4 9 4040 L 9 4 4 44 4 904440441400 4 44044080 T Sample size 128x128 80 Overlap Sample size 96x96 80 Overlap gt only half the dots detected gt only about 70 of the dots are detected 1 rt 4 F 4 ere r ze Ze Ze ze zu ze EPERE A O EEE AS IS gt 4 T 4 T 0 064 0F40644404 4 eo ee or 4 4 4 ren rt 4 4 CCC 0 er tt 44 4 44 TT Lt t tL t L L HH 4 4 tt 4 4 4 4 4400 00 4 4 4 44 OOO 444 ttt Tre eRe ee Sample size 64x64 80 Overlap Sample size 48x48 80 Overlap gt all dots found ideal parameter setting gt some dots detected twice resolved by increase of merging radius Calculation of the mapping function Once the grid is reconstructed the image mapping function may be calculated The mappi
76. ively P Vview can be installed from the setup CD ROM in several ways Installation Method A 1 Put the P Vview CD ROM into your CD ROM drive 2 In your file browser double click the file autorun exe This starts an installation option dialog 3 From there the setup procedure can be initiated Alternatively the documentation or latest release notes can be viewed 4 Follow the instructions in the installation program Installation Method B 1 Put the P Vview CD ROM into your CD ROM drive 2 From the Windows Start button click Run 3 Type the following Run command line D PIVVIEW2 SETUP SETUP EXE If necessary substitute the appropriate drive letter for your CD ROM drive in the above command 4 Follow the instructions in the installation program What is installed where By default the P Vview program and its associated files are installed in the systems Programs Of Program Files directory Other destinations can be chosen during setup All files associated with P Vview are copied into this directory tree that is no files are copied to the Windows system directories This also means that the program can be removed efficiently by deleting this directory although the systems uninstall method should be preferred Aside from the files the installation programs installs a few short cuts to the system s start menu PIVview makes extensive use of the Windows registry to store its current configuration settings These ar
77. led information about any calculated vector within the data set The vector closest to the mouse cursor is selected PIV gt Statistics presents a summary of statistics of the computed data set including the number of data points number of outliers various mean values and standard deviations as well as the maxima and minima Plot gt Plot Config opens a tabbed menu with vector and contour plotting parameters Short cut key p Plot gt Subtract Mean subtracts the mean displacement from the data set before plotting it Plot gt Print sends the currently displayed data set to the printer PlVview User Manual 11 Plot gt Position Objects allows free position of workspace objects such as lookup tables image or reference vector The PlVview Dialogs The Image Pre Processing Menu The purpose of this menu is to perform a variety of operations on the PIV image s prior to the PIV analysis This includes basic image re orientation image enhancement using filters background subtraction or mask overlays Each of these are described in more detail below Image PreProcessing El ES Image PreProceszing General Filters Background Mask General Filters Background Mask T Always Split PEO B16 Images IZ Convert Images to Floating Point Data Rotation i Inver for scanned negatives Y None T Median Filter 90 Left lt gt Right Bass size 33 i 180 C Top lt gt Bottom
78. low the user to view the stereo geometry from three major orthogonal directions The Apply button is used to transfer the settings to P Vview such that the reconstruction of the 3 C velocity map is repeated using the new settings StereoGeom may also be used in stand alone fashion In this case a file menu allows the user to load or save existing stereo project files Is ett ak Camera 4 a nn z aooo Camera B x FOO yfo 2 1000 Image size mm width E Height 200 fa Ref A uler StereoCfg module showing XZ projection of the camera light sheet geometry In this example the cameras are placed on either side of the light sheet PlVview User Manual 59 PIVmapT01 Projection Module This module is a semi automated procedure for determining the image back projection coefficients that are necessary for stereo PIV processing For input it requires images of a calibration grid see notes below for each camera view The grid reconstruction process can be split up into three major steps the coefficient calculation tab dialog follows these steps Step1 Line detection is performed by interrogating image locally using the INPUT Hough transform which transforms the image intensity distribution into a polar coordinate space Since straight lines show up as distinct bright spots in this transformation line segments are DEE eee ee CALIBRATION GRIDS Camera A Camera B
79. mage The larger the zoom factor the larger the resulting image Position objects puts PlVview into and edit mode where workspace objects image luts reference vector can be freely placed to suit the user s preferences Save current LUT Load LUT save or load user specified pseudo color lookup table Default positions restores the original positions of the workspace objects Save Load plot configuration stores restores plot parameters The Window menu Hide Toolbars hides all toolbars Individual toolbars can be activated or deactivated one by one The Help menu PlVview options displays the command line options available for PlVview About PivView gives information about the program itself e g release version number as well as shows the currently registered end user 10 PIVview User Manual The PlVview Toolbars File gt Open Image is used to load a new image for PIV processing File gt Open Data is used to load a new data set for post processing File gt Save Data Set opens a dialog with output options before saving the processed data set File gt Create Tecplot Data to save the processed data set in a format compatible with TecPlot Style files for TecPlot are also generated Image gt PreProcess opens a menu with image pre processing options Short cut key CTRL Image gt Edit Mask starts the mask editor program PlVmask Image gt Zoom out Image gt Zoom in increase
80. ng is a combination of two separate steps as shown in the figure below Mapping via polynomial Linear transformation using or ratio of polynomials magnification and offset GA gt Image space Mapped image space Object space X 5 y Xm Yen physical coordinates Xo Yo Zo The first mapping serves to de warp the image to reconstruct the image space onto a Cartesian grid At this point the image axes are parallel to the axes in physical space The second step merely applies a linear transformation to convert from image space to physical space Mapping from image recording to reconstructed image A variety of mapping functions are available for the first mapping step gt polynomials up to degree 4 48 PIVview User Manual X C x0 Te CX T Cu ya 2 2 T C 34 T C 4X Yi C 5 Y 3 d 2 af 2 2 C 6 CK yi Ct ys C x9 ye 4 3 2 3 4 CaoXi FC Xi Yi Cy Xj Yi TF Cy Xi PC ay Vi CPC Say 2 2 HE Cg VC ey 3 2 2 2 SE Cy6 T C 7 de Cak Yi C 9 Yj x 4 E S A 4 Coi T Oy Yi T Oyo Yi FH Y LA Y gt projection equations a ratio of polynomials 1 1 1 274 20 1 2m 279 _ dut F Ag Y tao F da dah dus Y ank Az Y lg dY A35X Y a y SE Ea Fa E Y T SOK e T a aX a35 X Y aY with a 1 The choice of function depends on the degree of distortion in the image In general if the grid consists of straight lines a ratio of first order polynomials should be favored However convergence
81. nnel readout etc The correlation plane window should be used in diagnosing the function s performance Multiple Correlation This algorithm essentially calculates the correlation from at least two slightly spatially separated correlation planes An approach based on multiplication of correlation planes was first proposed by Hart 2000 By calculating the product or sum of these correlation planes a new correlation plane results which ideally contains only the single displacement peak Due to the random position of the noise peaks which are present in all standard correlation planes the final correlation plane will show no or strongly reduced noise peaks Hence this procedure effectively increases the detectability of the correlation peak associated with the particle shift The algorithm requires additional processing time depending on the number of additional correlations calculated and does not perform equally well on all images A careful adjustment of the offset between the sample positions is sometimes necessary If correlation planes are too far displaced from each other the correlation signal may be lost entirely in high gradient flows The correlation plane window should be used in diagnosing the function s performance When the number of repeated correlations is higher more than 3 then the averaging mode should be preferred to the multiplication because this approach is less susceptible to loss of correlation in one of the repe
82. nstruction By manually moving these points the grid may be properly reconstructed nonetheless However these movements are lost if the scan operation is repeated Scan performs the feature extraction scan by the localized template matching scheme using the current parameter settings Any manual change to previously detected feature points are lost Build grid tries to reconstruct the grid using the currently detected feature points PlVview User Manual 52 Scale Tab Dot Spacing AK Control B specifies the actual spacing of the grid points This Fiters Scan Scale Map value is set in the main input menu of the program and Dat Spacing mm cannot be changed here SpacingX E Global physical offset Spacing Y assigns the coordinates of the calibration origin in llas physical space Initially this value is zero and should be set according to the experiment s reference frame PIN Camera position reflects the camera s position with respect to the origin J of the grid in the specified reference plane set in main sue ad menu These values are irrespective of the global o physical offset 000 0378527 Estimate Z mm 220 842 Underlying camera model tries to calculate the camera position using at least two calibrated planes A valid image mapping see Map is EEE El Camera pixel size um required The result even for non convergence is placed in the edit boxes above E E Various estimation schemes are
83. om the vectors This value is given in pixel displacement but is also displayed in physical units for reference This field only become active if the Subtract Mean option above is disabled Color Coding Magnitude color codes the vectors according to their magnitude Generally a rainbow type of LUT is used several are available The minimum and maximum magnitudes for the color bar can be specified in the fields below Signal to noise Ratio color codes the vectors according to their signal to noise ratio in the displacement estimation Validation color codes the vectors according their status of validation The color coding can be set in the Fonts Colors tab of the menu Uniform Color alternatively the vector can be plotted in a desired uniform color that is without color coding The color may be chosen through a color selection menu Apply Button any change to the setup can be immediately displayed by pressing this button Vector Style This sub menu provides further choices on the vector shape size and positioning ea Head width EE Head Length 34 a e Fill Arrow Heads F Uniform Heads Size le Uniform vectors Length f U IZ Anchor Yector at MidPoint PlVview User Manual 27 Contour Plotting Parameters Within P Vview a contour image describes a color coded representation of a scalar quantity derived from the PIV data These include velocity magnitude vorticitiy strain or single
84. ommand line program mapimage can also be used for this purpose Note that P Vview does not allow you to load images as long as the mapping coefficients are unity e g not calculated The back projected PIV images can now be processed by standard 2 C PIV interrogation In PlVview3C the same interrogation parameters are used for both views to ensure that the resulting data sets exactly match in size and position Command line based PIV processing is possible by the programs dpiv and dpiv3C The resulting pair of 2 C PIV data sets or files is now merged into a single stereo 3 C data set by taking into account the stereo viewing geometry The geometry information can be supplied relatively easily using the StereoGeom module Views of the geometrical PlVview User Manual 41 arrangement are available for three different orthogonal directions as a form of visual verification Data set merging is either performed by P Vview internally or by the command line program stereo In order to make file handling more user friendly all information relevant for stereo PIV processing is contained in a common project file This file is shared between the modules Individual modules only update their specific sections Platform independence is guaranteed through the use of the netCDF data format that is also used for the data sets as well as other parameter files Vector reconstruction in stereo PIV As mentioned above image pairs are first de warped and p
85. on classical and nonlinear peak detection image pre processing various filters masking region of interest selection interactive calibration outlier detection and replacement data postprocessing vorticity strain vector and contour display print capability data export TecPlot ASCII and netCDF PIV data import for further post processing WYSIWYG export using BMP PNG and others batch processing of image and data sequences data set statistics calculation stereo PIV processing capabilities drag and drop capabilities mask editing command line programs Since version 2 0 P Vview s user interface has been making use of the platform independent QT library Trolltech www trolltech com and can be made available on most Unix systems e g Linux Solaris Upcoming features Since P Vview is a continuously evolving program suite new features and improvements are constantly added PIVTEC is always open to new suggestions and encourages users to forward these as well as any bug reports What PlVview is not PIVview does not include any form of data acquisition such as PIV camera and laser control experiment synchronization etc This guarantees that the program can be operated on any PC system running a 32 bit Windows operating system Also P lVview has no intention of replacing the capabilities of full featured plotting packages such as TecPlot or Origin The plotting capabilities of P Vview merely provide the user with visual
86. orithm later snaps to the nearest grid crossing Step 5 Camera Positions This dialog allows the user to enter the positions of the camera with respect to the origin In addition the physical offset coordinates of the reference point may be specified Alternatively the StereoGeom program can be used to enter the coordinates Step 6 Save Project Finally the project must be given a name and is saved at the chosen location To begin with the actual grid detection start the P Vmap module by pressing the button The next steps show how the grid is recovered and the projection coefficients are determined Once the wizard is closed P Vview asks the user whether to load the new project file into PlVview3C Parameters such as the camera positions may be modified at a later stage from within PlVview3C Camera positions Help save Project Help Create new Stereo PIW Project Set Reference Point Set the reference point by dragging cross hair with mouse The further processing will lock onto the nearest grid crossing Create new 5tereo PIY Project The stereoscopic calculation requires that the position of the cameras is known With respect to a known reference point in the calibration grid This reference should preferably be visible in the grid image pair The coordinate convention is such that the plane coincides with the light sheet The asis Iz
87. orrection correlation analysis for both view particle image shift extraction data validation merging of the two views Y VV V V WV conversion to velocity The result is stored in a netCDF file such that it can be read by other PIVview programs Use ncdf2tec to convert the data to a Tecplot readable format Unless specified with the o option the output file name is determined from the first image name View A Calling procedure dpiv3c lt cfg_file gt lt imageAl gt lt imageA2 gt lt imageBl gt lt imageB2 gt dpiv3c pco lt cfg_file gt lt imageA gt lt imageB gt Parameter description lt cfg_file gt input P Vview3C generated stereo configuration file netCDF format lt imageA gt input PIV image pair for view A lt imageB gt input PIV image pair for view B Command line options PIVview User Manual 69 align perform camera disparity correction def performs second order accurate image deformation PIV mg lt fact gt performs multi grid pyramid PIV initial de sampling factor lt fact gt default 3 info provides feedback during processing of data o lt fname gt stores output data in file lt fname gt noip avoids image pre processing options pco treats the input images as a PCO double exposure image both images must be PCO images wWX y specifies the interrogation window size Overrides the value specified in the parameter file 5x y specifies the interrogation window step size Over
88. pixel and 1 bit pixel including compression Not supported are LZW compressed TIFF images on which Unisys holds a patent If you plan to use LZW TIFF images use a image converter such as IrfanView to convert the images to another supported format Color images are converted to 8bit grayscale Windows Bitmap BMP 1 4 and 8 bit pixel color images are converted to 8 bit grayscale JPEG Color JPEG s are converted to 8 bit grayscale Portable Network Graphics PNG Color images are converted to 8 bit grayscale Portable Bitmap Graphics PBM PGM PPM This includes various formats from a collection of ASCII based bitmap types PBM PGM and PPM PBM Portable Bitmap format is a 1 bit black and white bitmap type PGM Portable Greymap format is an 8 bit greyscale bitmap type and PPM Portable Pixelmap format is a 24 bit full colour bitmap PCO B16 a 16 bit pixel image format used by PCO s camera acquisition software B32 an extension of PCO s B16 format to 32 bit pixel LaVision Images IMG a 16 bit pixel integer and 32 bit pixel floating point image format used by LaVision s DAVIS software Not supported is LaVision s proprietary compressed IMX file format Note The command line utilities such as dpiv only support the following image formats gt TIFF all except LZW TIFF and 24 bit color gt BMP 1 and 8 bit pixel uncompressed gt PCO B16 gt LaVision IMG Image naming and numbering conventions Generally PlVv
89. ptionally replaced with new data Two replacement schemes are available Other Correlation Peaks The program stores the three highest correlation peaks found in each correlation calculation If the highest correlation peak turns out to be an outlier this options picks the next highest correlation peak and repeats the outlier detection procedure Interpolation Applies a bi linear interpolation to estimate the replacement vector When several of the immediate neighbors are also outliers then a Gaussian weighted interpolation scheme is used Re evaluate with larger sample If none of the alternative correlation peaks fulfill the validation criteria the image can be locally re evaluated with the next largest interrogation sample in the multi grid processing scheme Note All correlation peaks are first tried out before any data is interpolated The Apply Button can be used to immediately see the effect of the outlier detection and replacement provided that the image is already evaluated 22 PIVview User Manual PIV Evaluation Setup Conversion This dialog is used for setting the magnitude and time conversion factors which are necessary to obtain velocity data in a physical grid Within P Vview the horizontal display axis is always associated with the X axis the vertical with the Y axis No Z coordinate is used If the measured PIV plane is not aligned with the desired physical space coordinates the user has to perform the required
90. ra B if loaded Show combined data displays the reconstructed 3 C data if available PIVview User Manual 43 Creating a new stereo project with PlVmap3 The menu entry Stereo lmage Registration calls up the program PIVmap3 which allows a step by step calibration of a stereo PIV configuration including the grid registration procedure Background on Stereo PIV Most planar imaging techniques require some sort of calibration to associate image coordinates with physical coordinates A viewing arrangement in which the object plane e g laser light sheet is parallel to the image plane co planar arrangement involves a simple linear transformation consisting of translation scaling and possibly rotation This is no longer the case for oblique viewing arrangements or if lens distortion is present In effect the co planar on axis viewing arrangement should be considered as a special case of the oblique viewing arrangement The calibration procedure for the more generalized oblique viewing arrangement typically requires high order mapping functions to transform image pixels to physical space coordinates Another calibration step is necessary for multi component measurements such as DGV or stereoscopic PIV In this case the positions of the cameras with respect to the image plane are required The most pragmatic approach is a direct measurement of the camera positions with respect to the object plane Willert 1997 which is not necessarily trivi
91. rams cccccccesseeeseesseeeseeesseeseeesseeseeanseeseoenseessosseeseonnneesees 67 O A ee eee ee 67 EIT 67 nn A y 68 IMAGE PRB 2 522 anni 70 NSBFZTE Gr thse ee een 70 NOVOME sense ee re A 71 NOGEN Hesse aaa else 71 POS PRO ee 71 SA e ONT Eg DUEL 72 Sh rte t keys IN UA TTT 72 Mouse ShortGuis amp K eT 73 VI PIVview User Manual Supported Fil Formats TT 74 Supported MAGS Formal en een 74 Image naming and numbering conventions u uu044400nnnnnenn nenne nenne nenne nenne nenne nenne nennen 74 supported Output File Formats ati een 75 References and Further Reading 2a20 0200na000nnan0nnnnunnanunnnnnunnanunnnnnnnnnnennnnunnanenn 75 Overview of PIVview Background PIVview is a compact program package for the evaluation of particle image velocimetry PIV or speckle displacement recordings Its original and primary intention was to provide a quick look data retrieval from PIV recordings Experience has shown that this quick look capability is essential in the optimization of the acquisition process Features During its development P Vview has matured from a simple PIV image probing program to a nearly complete PIV evaluation package including e advanced processing algorithms multiple pass method mulitple grid pyramid method double correlation method second order image deformation methods various peak detectors and fitting algorithms ensemble correlati
92. ricted PI fig Advanced PIY Processing Options Store intermediate deformed images Multigrid Final scan File prefix deform Iterations B 3 7 Sample overlap for multigrid processing M Limit peak search area None 25 d BUS C 75 Em 2 on final pass only 18 PIVview User Manual PIV Evaluation Setup Peak Detection g PIY Evaluation Setup Grid Correlation Algorithm Peak Detection Validation Conversion Filters etc eee Pixel Peak Fit Nor lnesr Gaussian Nor inear Gaussian peak ft lt fit Peak Fit size E 4 Limit Peak Search Area Horiz 5 py ES Vert 5 py Distance from nami Disable Self Correlation Peak Maximum distance tron 00 4 0 Apply Cancel Sub Pixel Peak Fit Various peak fit algorithms are available which recover the sub pixel displacement of the correlation peak 3 Point Gauss Fit This is the most frequently used scheme as proposed by Willert amp Gharib 1991 It selects the four closest neighbors of a correlation maximum and fits a 3 point Gaussian curve along each of the two major axis Least Squares Gauss Fit Similar to the 3 point Gauss fit but uses all 8 points surrounding the maximum Since it uses more data for the location of the peak the noise in the result is generally less than with more simple fits This peak fit method is recommended for most applications 3 Point Parabolic Fit Similar to the 3 point Gauss fit except tha
93. rides the value specified in the parameter file bx y specifies the global image offset Overrides the value specified in mask lt fname gt uses specified mask image Useful for batch processing with varying masks Note mask image must have the same size as PIV images 70 PIVview User Manual IMAGEPRE This utility program reads the specified image and performs image pre processing filtering background subtraction etc as specified in the parameter file Calling procedure imagepre lt par_file gt lt imageIN gt lt image0UT gt Parameter description lt par_file gt input PlVview generated parameter or data file netCDF format with image pre processing information lt imageIN gt input original image lt imageOUT gt output pre processed output image Command line options nO La avoids flipping or rotating the image important for stereo PIV processing norotate avoids flipping or rotating the image important for stereo PIV processing FLATO provides feedback during processing of data PEO treats the input image as a PCO double exposure image and generates two output images with the extensions ima imb ea STT same as the pco option but stores the resulting image pair in a single PCO double expsure image file NCDF2TEC The program reads the specified P Vview data file and generates a Tecplot formatted ASCII or binary output file Calling procedure Hedrzree lt tileTN gt lt rLLe uUT gt Paramet
94. rocessed for the respective view using standard 2 C PIV correlation processing algorithms The resulting 2 C PIV data points from both views lie on a common grid but have to be combined using the local viewing angles The figure below illustrates the viewing geometry in the xz plane y X ZAS w 1 0 oO 7 O K o 1 2 Pa Ay O P U Ax li o 7 P w Y X Ay Z Ax Ach o1 T Z P O a P L AX AV V ATAJTATAX The system of equations given on the left comprise a once over determined 2 unknown velocity components vs 4 known displacements and are solved in a least squares fashion The residuals of this solution are a measure of quality of the vector reconstruction In practice the residuals should be below 0 5 pixel The validation tab of the PIV configuration menu allows a vector validation based on these residuals Enabling the Stereo Mode The stereo modules of P View are enabled by selecting the menu entry File Stereo A check mark next to the option signals that P Vview is in stereo PIV mode The Stereo menu is then enabled Once the stereo mode is enabled certain features of standard 2 C P Vview are disabled gt The calibration mode is no longer available All calibration is now supplied through the PIVmap image registration module that is the physical grid spacing together with the mapping coefficients define the image magnification factor The origin is also defined through P Vmap gt PIV images
95. rocessing step may use faster peak fitting algorithms or simpler image interpolation schemes This may have a minor influence on the final result e g reduced precision and or reduced vector detection rate load image pair and parameters set up initial grid standard PIV analysis cross correlation A PIV Image VA yA VA 14 N X x 7 j j N a Pa N a Pa a N Na N N N N aw a N N N save data Yi ENT N N E ENT N N CL Flowchart for coarse to fine multi grid processing with image deformation for second order accurate PIV interrogation PlVview User Manual 17 Advanced PIV Processing options Store intermediate deformed images is active during image deformation schemes If checked the program stores the iteratively deformed images for each intermediate processing step after image deformation Sample overlap for multi grid processing determines the sampling grid for the intermediate steps of mutli grid schemes Multi grid Final scan determines the number of repeated interrogations at the final grid resolution Generally one final scan is sufficient whereas additional scans may increase the precision Limit peak search area is active only for the final passes of multi grid algorithm at the final grid resolution In general the iterative schemes should converge on the underlying data within a few pixels or less such that the correlation peak search area can be rest
96. rtical bar on the left in this case a linear intensity ramp with O black at the bottom and 255 white at the top The number at the top of the LUT represents the intensity of brightest display pixel The image brightness can be changed by using the menu entries Image Brigher or Image Darker or the short cut keys b and d respectively Pixel intensities exceeding the maximum display intensity are either clipped at the maximum or wrap back to the bottom of the LUT Image Clipping The display of the LUT can be turned off with the menu option Image Hide LUT Main Window Reference Vector The reference vector s main purpose is to have a visual reference when a complete data set is plotted or exported It s magnitude and display is defined via Plot PlotConfig Ref Vector Main Window Scroll Bars The scroll bars on the right side and bottom edge of the main window can be used to pan the image to a different viewport if the entire image does not fit into the display window Status Bar The status bar generally reports the cursor position in both image units pixels as well as physical coordinates Also the image intensity is given In case of interactive interrogation the correlation result is shown e g position displacement velocity correlation coefficient If the image has been evaluated these fields provide information about the data point closest to the cursor Note Many of the toolbar buttons and menu entries only becom
97. s or reduces the magnification of the image on the screen If the image no longer fits the screen scroll bars on the edges allow image panning Short cut keys Image gt Brighter increases the image brightness on the screen Short cut key B Image gt Darker decreases the image brightness on the screen Short cut key D Image gt Auto scale sets default intensity on image Image gt Change LUT changes the display look up table LUT Each click of the icon picks the next or previous LUT in a circular list When scalar data e g vorticity is displayed in place of the PIV image the accompanying LUT is changed Short cut key L PIV gt Calibration opens a menu for online length calibration Only the image with an overlaid ruler and origin are displayed Short cut key F3 PIV gt PIV ROI changes into selection mode region of interest ROI Short cut key F2 PIV gt PIV Configuration opens a tabbed menu with PIV processing parameters Short cut key F4 PIV gt Evaluate Image performs evaluation of PIV image using current configuration defined by the PIV processing parameters Short cut key F5 PIV gt Batch Processing opens the batch processing menu PIV gt Correlation Plane displays the correlation plane during interactive interrogation using the mouse PIV gt 2 D Histogram displays a global histogram of the currently computed data set PIV gt Data Probe shows a window with detai
98. s the creation of irregular polygon shapes Points are added to the polygon with the left mouse button The right button closes the current polygon Draw Delete object deletes the currently selected object PivMask C DPI p1 ima lol x PivMask C DPI p1 ima File view Draw Help File view Draw Help Eer ARER iO K X RAK x Left mouse button to add vertices right button to Finish 7 Left mouse button to add vertices right button to Finish PlVmask in mask display mode Objects PIVmask in mask display mode showing extending beyond the image dimensions an inverted mask have no influence on the mask 40 PlVview User Manual Stereo PIV Modules Stereo PIV processing is possible through a separate set of modules for P Vview These modules are tied into P Vview3C but can be also be used in stand alone fashion Stereoscopic PIV is considerably more complex than standard 2 C PIV implementations The use of the stereo modules is not quite as straight forward and requires that the user has a basic understanding of stereo PIV implementations namely the concept of image registration through back projection Aside from reading this brief introduction new users are encouraged to study the stereo PIV example s supplied on the program CD Stereo PIV Procedure in Pil Vview3C The figure at left illustrates the general Camera A Camera B procedure followed by PlVview for stereo PIV Image Pair PIV Image Pair PIV processing
99. seem to account for distortions introduced by tilting of the back plane Scheimpflug angle adjustment in stereoscopic PIV In addition to the estimated camera position the Tsai camera model also yields distortion coefficients and estimated lens focal length which are not necessarily of direct interest in image coordinate calibration A slightly different approach to camera calibration is described in the following and has been implemented in the P Vmap3 program The basic aim is to provide an accurate image mapping to de warp images on the hand as well as to retrieve the position for each individual camera without knowledge of the camera characteristics as in the pinhole camera model i e pixel pitch lens focal length The image mapping function is calculated from a single recording of a calibration target that is placed in the object plane The target should contain a known pattern such as a rectangular grid of lines dots or crosses as shown in figure 4 Calibration of the camera position is also based on an image series of the Z translated targets as proposed by Soloff et al 1997 The approach followed here is to use a functional relationship between the out of plane displacement and associated in plane displacements to estimate the camera position In practice this involves solving the projection equations for 44 PIVview User Manual the point at which the image coordinates x y vanish With the object coordinates xs yo Zo and image
100. sked pixels experimental Tab Background Subtract Background When activated the specified image or image pair is subtracted from all data images upon loading Background subtraction can improve the data yield in images suffering from strong background artifacts The image size of this background has to match that of the PIV image Separate background images for each image of a pair may be specified Tab Mask Use Mask When specified mask images are used to blot out data in regions covered by the mask Separate mask images for each image of a pair may be specified Invert Mask Inverts the pixels of the mask images prior to application Randomize masked pixels This is an empirical method to reduce the influence of the masked areas on the correlation signal Normally masked pixels are set to zero prior to correlation This may result in edge artifacts in the correlation plane In this case the masked area Is filled with random data that should have roughly the spectral characteristics of the unmasked data and generally reduces the edge artifacts in the correlation plane PlVview User Manual 13 PIV Evaluation Setup PIV Evaluation Setup Grid This dialog can only be opened when an image pair is loaded The image is then overlaid with a grid of crosses indicating the centers of the correlation samples i e interrogation windows g PIY Evaluation Setup EH Horizontal Vertical Grid alignment Bottom edge r Window Si
101. splayed The overall speed of the mapping process depends on the output image size and the number of reconstructed planes PlIWview User Manual 53 Helper axis Reconstruction of the grid is based on the a priori knowledge of the direction of the major axes as well aS an approximate location of the origin This is achieved by manually specifying a pair of helper axes for each calibration image The figures below show how these axes may be placed in the image using the mouse The tips of each axis should be placed on the points neighboring the origin to allow a more efficient grid reconstruction The function Copy helper axis can be used to copy the coordinates from the previous plane to the currently displayed plane This speeds up positioning over many calibration images Position of axes at beginning Dragging the X axis to the origin Alignment of Y axis and arrow tips Figure 6 Manual positioning of the helper axis in the image Description of the P Vmap3 menu entries File New creates a new PlVmap3 project All previously loaded images are removed and file names are reset Sy File Open loads a previously created PIVmap3 project All current settings are lost File Save stores the current selection of image and parameters in a user specified project file This ASCII file is editable with any text editor i e Notepad The default extension is map n File Save stores the current selection of image and paramet
102. t it is only fits a parabola through the three points Center of Mass Fit This sub pixel locator uses a specified number of correlation values surrounding the maximum 3x3 5x5 etc and calculates the associated first moment or center of mass It is generally not well suited for the narrow correlation peaks recovered from typical PIV images In cases where larger particle images or larger features have to be correlated the correlation peak will be wider and this fitting algorithm may yield more accurate data Whittaker Reconstruction Recnstructs the correlation function from all correlation values along the horizontal and vertical line crossing the brightest correlation value peak and estimates the maximum from this function Nonlinear Gaussian Peak Fit Uses up to N points for a parametric fit to a Gaussian shaped function allowing for peak ellipticity The peak shape is estimated from the autocorrelation function of the image samples which may reduce overall processing speed Limited Peak Search Area When activated the correlation peak detection is restricted to an area whose size can be selected For data with a small dynamic range a limited search area can significantly increase the data yield That is only the area enclosed by the given rectangle is scanned for the correlation peaks Disable Self Correlation Peak PIVview User Manual 19 disables the use of the correlation peak near the origin which may be useful in images with
103. th TecPlot netCDF is a self describing platform independent data file format frequently used in the scientific community NetCDF data generated by PlVview is aimed to be compatible with the layout utilized by PivNet EuroPIV cooperation projects Since NetCDF is an open format additional variables can be added without interfering with a predefined layout Further information on NetCDF can be found at Unidata www unidata ucar edu Two separate utilities are redistributed with PivView that allow NetCDF file viewing ncdump as well as NetCDF file generation ncgen In principle these utilities allow the user to edit a NetCDF file by first converting it to a text file with ncdump and recompiling it after modification using ncgen Parameter Files The current parameter settings are generally stored in the system registry when the program exits Once the program restarts these parameters are restored However the most important parameters especially those regarding evaluation conversion and post processing can also be stored in a separate file which has the same name as the image and the extension par These files are formatted in NetCDF see above and are compatible with the existing UNIX based PIV software of the German Aerospace Center DLR as well as with the specifications defined by the EuroPIV PivNet activities References and Further Reading Raffel M Willert C amp Kompenhans J 1998 Particle Image Velocimetry A Pra
104. the grid may be locked to any specified point pixel the image Use ROI Restricts the evaluation to a defined rectangular region within the image This region of interest ROI can either be set by entering boundary positions or interactively by using the ROI button on the toolbar or menu entry PIV gt Select ROI Process Button closes this menu and starts evaluating the image 14 PIVview User Manual Default Button resets all values in this menu to their default values PIV Evaluation Setup Correlation Eu PIY Evaluation Setup T Phase correlation Phase only Filtering T Nyquist Frequency Filtering Y Multiple repeated Correlation Number of repeated correlation la 4 Image r sampling C correlation re sampling based on grid spacing A 16 px Y 16 px Y user specified correlation re sampling offset x d px Y jo px Correlation accumulation mode C Multiplication of correlation planes f Averaging of correlation planes Phase correlation Phase only correlation This approach spectrally filters the data such that only the phase information is used in the calculation The technique proposed by Wernet 2005 has the potential of obtaining a higher correlation signal Nyquist Frequency Filtering This option removes the highest frequency components in the cross power spectrum Fourier transform of the correlation function This can be of use in the removal of camera noise artifacts interlacing multi cha
105. um or maximum cut off thresholds by summing up the individual bin counts of the image histogram The threshold is value is determined when the 12 PIVview User Manual percentage of pixels in the sum exceeds the percentage specified by the user Obviously this only works for images which a not binary High Pass Filter performs a high pass filtering of the image by subtracting a lowpass filtered version of the input image Both the kernel size and filter shape Rectangle or Gaussian of the low pass filter can be specified Gaussian filtering somewhat slower but positively influences the final result Negative values are clipped only if this option is activated Binarization binarizes the the image data by setting pixel values above the specified threshold to white 255 and lower values to black 0 AntiAlias Smoothing applies a low pass filter which can be used to damp out spurious noise in the image Tip The combination of the last three filters can significantly help in processing images with low signal to noise ratio Image PreProcessing EIE Image PreProcessing EIE General Filters Background Mask General Filters Background Mask M Subtract Background Choose WM Use Maskl C PI PIV_Challenge2005 B min_ima ti Choose Mask MT Use seperate bkod for frame B Mb Choose IE Pros ampleD ataReal_BLAWallM ask bmp SSeS See Use separate mask for image E Ehoose Maske MT Invert Mask s T Randomize ma
106. ut last file in sequences to be processed Command line options 72 PIVview User Manual prefix lt stub gt overrides the output file stub to the name provided A use parameters associated with camera view A if available B use parameters associated with camera view B if available No provides feedback during processing of data tec store result in Tecplot ASCII file SOLL store result in Tecplot binary file STEREO The program reads the specified PIVview data files supplied for each camera view and reconstructs the 3 C velocity field using the geometry information specified in the parameter file Use the StereoGeom module to specify the geometry information Calling procedure stereo lt CfgFile gt lt fileA gt lt fileB gt lt fileOut gt Parameter Description lt CfgFile gt input Stereo project file as generated by P Vview3C and or stereo modules lt fileA gt input PIV data set for view A netCDF format required lt fileB gt input PIV data set for view B netCDF format required lt fileOut gt output PIV data with reconstructed 3 C velocity data default netCDF format Command line options info option gives feedback during processing plt option writes out the data in Tecplot binary format tec option writes out the data in Tecplot ASCII format A use grid from lt fileA gt as output grid If the grid from view B does not coincide then this data is interpolated B use grid from lt fileB gt
107. with by changing the parameters in the Projection tab of the parameters Tools Eval parameters opens a tab dialog that allows step by step processing of the grid images Tools Show Info gives information about the status of P Vmap Tools Measure places a ruler into the image allows the user to measure distances between two points Tools Set ROI allows interactive definition of a region of interest in which grid detection is to be performed Tools Elliptic ROI toggles the ROI shape between rectangular and elliptic Tools Best fit ROI maximizes the ROI to fit the image size View Show Grid A displays the calibration image for camera view A View Show Grid B displays the calibration image for camera view B View Back projected image s dewarps the loaded calibration images using the current sets of projection coefficients if avaiable The View menu has a number of entries that are identical to those found in PlVview zoom functionality image brightness etc PIVview User Manual 61 gt PivMapTOL C DPI stereotest myStereo corr chg 5 xj File Tools View Help ISS d 3 SERERE la B Qm HO Y Main window of PIVmapT01 showing calibration image overlaid with reconstructed grid Red crossings indicate the major axis The origin itself is marked by the yellow circle Note that the image is a negative of the original which consists of dark lines on a bright background PIVmapT01 Tab Dialog
108. ze px 32 r 32 r Grid Anchor Point Step Size px E 16 pe 24 a pz 24 Overlap 50 50 Image Offset px fo fo Current land Dimensions 16 1424 gt 89 Nodes Tel 16 672 gt 42 Nodes gt 3738 Nodes total Use ROI Fit ta mane Window Size The menu entry defines the interrogation window size that is the rectangular size with which the PIV images are sampled to perform the local cross correlation analysis Frequently the interrogation window Is also referred to as interrogation spot The window dimensions can be freely set e g 24x27 is also possible Step Size defines the increment with which the images are sampled These values define the mesh size of the final data size Typical values are around 50 of the window size which satisifies the sampling criterion Greater sample overlap e g smaller step sizes produces over sampled data but does not necessarily yield additional information Image Offset Applies a constant integer offset to the interrogation windows by offsetting the images with respect to each other This offset is applied to the displayed image with the exception of double exposure autocorrelation images Interrogation window offset is especially useful in the analysis of images with large convective displacement e g pipe flow wind tunnel flow etc Grid alignment Allow the grid to be aligned either with the bottom or top edge of the image Alternatively

User Manual - PIVTEC GmbH

Contents

Download Pdf Manuals

Related Search

Related Contents