GUI User Manual
Contents
1. Display Image Color Display Image Gray Figure 14 We focus the attention on longitude from 9 2 to 9 7 min and max LON and latitude from 43 4 to 44 1 degrees min and max LAT 21 COVIS 3 0 User Manual Figure 1 EDENES Fie Edt View Insert Tools Desktop Window Help Dee khaa 08 eo Figure 2 File Edit View Insert Tools Desktop Window Help Dee kaato A 08 so Selected area Covariance Result 0 04 44 laea 0 035 0 03 SF VA 85 Ye YT ES 0 025 16 18 20 22 2 2 002 5 Residual Residual s Histogram ils 5 2 9 0 005 fJ i i i i i o zi r Distance Km Figure 15 a Mean field and Gaussian Residual field of the selected area b Isotropic covariance is modelled by an exponential law The intrinsic standard deviation of the filed is 0 035 and the e folding distance around 7 km Figure 4 LILA Figure 3 File Edit View Insert Tools Desktop Window Help File Edit View Insert Tools Desktop Window Help Des BAAD 08 a DeSean Ea n Merged image Difference between original and Merged Image eK E E E IGS Eby 26 26 28 925 93 J 94A 945 95 J 96 965 97 Figure 16 Results from the merging the white dots represent the ship trajectory a Merged Image b difference between merged and original image 22 COVIS 3 0 User Manual 7 REFERENCES 1 G Pennucci A Alvarez and C Trees A Sa2. 44 2 44 1 44 43 9 43 8 43 7 43 6 43 5 43 4 43 3 2 2008 0820 1211 n18 sst gtif minLON insert min LON deg maxLON insert max LON deg min LAT insert min LAT deg mex LAT inset meckAT asa _ Display Sub area 8 6 8 8 9 9 2 9 4 9 6 9 8 10 10 2 10 4 Reset LAT LON Longitude deg 2 23 24 5 26 27 28 Reload New Sat Image To change the colormap insert here minimax values Display Image Color Dismay Image Gray Figure 13 Satellite and in situ data ingestion NURC GUI can read several formats of data as described in detail in Chapter 5 Once time the image and the in situ track have been ingested click on the button Zoom in to zoom on the area of interest and to identify the corners for the processing The Min Max values of the color scale are displayed automatically but they can be changed inserting the desired values on the lower part of the window In this case to better visualize shapes the color scale Load Satellite Image Load in situ data std_obs oor ta a io std_sat 05 Latitude deg ta o D t a on Covarianza Display Merged Image S n 0 o i o ny l Eoo 8 8 9 9 2 9 4 9 6 9 8 10 10 2 have been re defined between 22 and 28 degrees e COVIS3 0 2008 0820 1211 n18 sst gtif Reset LAT LON Longitude deg 22 23 24 25 26 27 28 To change the colormap insert here min max values 2
3. COVIS 3 0 User manual Revision 3 0 2012 April 1st NZ Prepared by G Pennucci A Alvarez and C Trees NURC NATO Research Centre 19100 La Spezia Viale San Bartolomeo 400 Italy Tel 39 0187 527 318 Fax 39 0187 527 354 Copyright 2012 All rights reserved Copyright COVIS 3 0 User Manual Copyright 2012 all rights reserved worldwide This user s guide is protected by federal copyright law No part of this user s manual bay be copied or distributed transmitted transcribed stored in a retrieval system or translated into any human or computer language in any form or by any means except by prior written permission of NATO Research Centre NURC Italy or Naval Research Laboratory NRL Stennis Space Center MS Technical Support Technical support for COVIS 3 0 can be obtained in accordance with the user s license agreement from Giuliana Pennucci Tel 39 0187 527 318 Fax 39 0187 527 354 pennucci nurc nato int If you encounter a problem during a COVIS 3 0 run please send an e mail with the MATLAB printout error along with a description of the problem This will greatly increase the speed at which we can help troubleshoot any problem 2 COVIS 3 0 User Manual TABLE OF CONTENTS INTRODUCTION 4 1 1 Ways in which COVIS 3 0 can be used 5 SYSTEM REQUIREMENTS 5 INSTALLING COVIS 3 0 6 OVERVIEW 6 4 1 Ways in which COVI
4. 0 this is possible because APS produces 12 COVIS 3 0 User Manual standard products and the geographic information have same name and same format for each output HDF List ee COVIS 3 0 Select one variable from the list below and click on the buttom to upload Normalized water leaving radiance at 488 nm mW cm2 umi sri bb_531_qaa a bb_547_qaa bb_667_qaa c_412_qaa min LON insert min LON deg r max LON insert max LON S ie ES ee aaa p a min LAT insert min LAT deg max LAT insert max LAT deg Ingest in COVIS i A i i i i a i i 2 425 19 Os G BS be BS Longitude deg User Selection To change the colormap insert here minimax values min val max val Display image Color oisney image Gray Figure 6 HDF data viewer for files generated with APS software The user has to select the desired product from the left list and then click on the button Display Image to visualize The user can chose the product to process repeating this procedure several time when an interesting image was identified he has to click on the button Ingest in COVIS 3 0 to start with the processing Once time an interesting product was identified the user can start with the data processing clicking on the button Ingest in COVIS 3 0 as explained in detail Chapter 6 To test this procedure we have provided three examples of HDF generate
5. R Arnone available in the SPIE public Journal COVIS 3 0 is currently supported by Naval Research Laboratory Stennis and developed at NATO research Centre NURC Italy The COVIS 3 0 source code is written entirely in MATALB in order to make it easily portable to almost any computer with a MATLAB compiler Throughout this report the names of mathematical variables are written in italics e g C or cov or covariance The names of computer programs directories and files are written in Courier font 4 COVIS 3 0 User Manual 1 1 Comments and Warnings It s important to understand that the COVIS 3 0 tool per se is an implementation of a statistical model therefore its results are directly related to the size and quality of the input parameters For example a limitation of COVIS 3 0 is the quality in terms of noise cloud coverage percentage of the satellite images used as input It is you job to select areas with low cloud coverage and to build consistent data set before using them as input to COVIS 3 0 for more details see Paragraph 5 2 COVIS 3 0 is reasonably robust in checking for bad input but there is still great opportunity for entering bas input and getting bad covariance output More details on this type of problem can be found in Reference 3 Another limitation of COVIS 3 0 is related to the memory of your machine s CPU and to the size of the selected working image for the covariance evaluation For a proper covari
6. User Manual 6 EXAMPLE COVIS 3 0 RUNS In this paragraph we report an example of COVIS 3 0 utilization in particular we present an example of implementation covariance analysis and merging that has been retrieved using SST AVHRR image and an in situ ship track that have been performed by the N RV Alliance NATO vessel STEP 1 Set the MATLAB current directory in the GUI folder path e g D Desktop COVIS as in Figure 12a To launch the program from the Command Window type covis in the MATLAB Command Window and press ENTER as in Figure 12b MATLAB 7 5 0 R2007b File Edit Debug Desktop Window Help DHa 4s MO SO curent Directo Desktop covis J e Shortcuts 4 How to Add 4 What s New Workspace oa x Sm Sa S EE Name value Fo see Demos or read Getting Started Figure 12 Initialization procedure to run COVIS 3 0 the user has to set own current directory and launch the program from the MATLAB Command Window STEP 2 From the COVIS GUI click on the button Load Satellite Image to ingest the SST AVHRR image D Desktop COVIS SAMPLE_IMAGE GEOTIFF 2008 0820 1211 n18 sst gtif following the instruction described in paragraph 5 1 and displayed Figure 2 STEP 3 To load the in situ data click on the button Load in situ data and select the in situ that are available here D Desktop COVIS sample_insitu_ligurian The data set used in this example represents a SST track performed by the N RV Allia
7. coordinate systems ellipsoids datums and everything else necessary to establish the exact spatial reference for the file It s important to note that COVIS 3 0 was written to read any GeoTIFF file that contains one georeferenced image that means a two dimensional of M by N pixels array Moreover to a proper covariance processing the info GeoTIFF structure have to contain the following field ImageDescription describing the image information such as 9 COVIS 3 0 User Manual units scaling slope and intercept This information is important to a proper signal processing because is required to convert data from numerical to physical If no description is included in the file the field is empty or without this information COVIS 3 0 will show the image as was stored in the file considering the data as physical supposing that the numerical to physical conversion have been performed by the TIFF GeoTIFF provider 5 1 2 MATLAB format mat This type of input format was introduced in COVIS 3 0 for expert MATLAB users that want to process directly MATLAB gridded data The input mat file has to contain the following parameters 1 Image a grid of M rows by N columns that represents the image of interest physical values 2 Latitude decimal degrees a vector of 1 row by M columns size has to be the same of the grid image 3 Longitude decimal degrees a vector of 1 row by N columns In particular when a mat f
8. gt is the observation error matrix The first part in the exponential represents likelihood density while the second product of the matrices 18 COVIS 3 0 User Manual represents a priori probability This merging procedure is performed maximizing the a posteriori probability distribution therefore the best estimation is represented by the field y Y merged arg miny Ay H Da Voss Ay v z yy Cc y T y 5 The solution of Equation 5 represents the merged image In the next Chapter we present some results on the merging procedure however before to proceed it is important to point out the fact that the merging technique can be applied to in situ measurements that verify the following conditions exclusion criteria In situ and satellite acquisitions have to be measured at the same day and same time or at least with 1 hour of difference In situ acquisitions have to be taken with low wind condition less than 12 ms and when the solar zenith angle is lesser than 70 degrees The satellite area has to be cloud free or with a cloud cover less than 10 15 This condition is important because it has been demonstrated that covariance analysis is directly dependent on clouds size position amount of information in the sense that the introduction of clouds causes a loss of information and decrease in the correlation length as expected More details can be found in References 2 and 3 19 COVIS 3 0
9. 43F min LON insert min LON deg std_sa 05 AZ von tneon maxLON insert max LON deg B 2 v 5 a Biibescdeace i ae gt i min LAT insert min LAT deg 40 3 i N maxLAT insert max LAT deg 39 i i i 8 10 12 Longitude deg 0 0 5 1 15 2 25 3 3 5 From here user can change scale and colormap To change the colormep insert here minimax values nin val max val Display image Color Disney WGA Figure 7 Chl visualization from a MERIS acquisition Using the color scale of the file that image is not well represented therefore the user can also decide to change the color representation using the lower buttons in the lower part of the GUI as in the black box epe COVIS30 099 Losdin stu dta soos oD T insert min LON deg stase 05 iwert max LON deg insert min LATI eg insert max LAT deg Display Sub orea 12 Reset LATLON Longitude deg ra Reload New Sat nage To change the colormag nsert here minimax values 08 Otsplay mage Color Oteplay image Gray 14 COVIS 3 0 User Manual min LON insert min LON deg LON insert max LON deg min LAT insert min LATIdeg AT insert max LAT deg Display Sub srea Figure 8 Chl visualization of the same a MERIS acquisition of Figure 7 but using user defined min and max values of the colormap If desired the user can decide to use gray scale color clickin
10. Higher resolution monitors could make the GUI form appear smaller than intended in which makes the forms harder to read 5 COVIS 3 0 User Manual 3 INSTALLING COVIS 3 0 This section describes how to install COVIS 3 0 on a PC It is assumed that you will be running COVIS 3 0 with a MATLAB compiler of Version 7 or more COVIS 3 0 is available as a standard MATLAB GUI as a single self program from your MATLAB Command Window COVIS 3 0 software package is distributed as one directory named COVIS 3 0 To install the code on your PC do the following steps l Open MATLAB and set the current directory in which COVIS 3 0 code will be installed COVIS 3 0 2 From the Command Window launch the program COVIS and the COVIS 3 0 GUI will appear 3 Follow the instruction on Chapter 5 to ingest and process data pushing the button on the screen 4 OVERVIEW OF COVIS 3 0 4 1 Ways in which COVIS 3 0 can be used This section gives brief description on how COVIS 3 0 can be used what assumptions are building into the code what input is required and what output can be obtained from a COVIS 3 0 run Some of the ways in which COVIS 3 0 can be used are as follows COVIS 3 0 can be run to ingest and visualize several remote sensing satellite images such as from SeaWiFS AVHRR MODIS MERIS VIIRS and all the satellite outputs that have a standard input e g HDF HDF4 HDF5 and Geotiff format COVIS 3 0 can be run with satellite remote sensing i
11. S 3 0 can be used 6 4 2 Inputs 6 4 3 Outputs 7 GRAPHICAL USER INTERFACE ORGANIZATION 8 5 1 Data Ingestion 8 5 1 1 TIF and GEOTIFF format 9 5 1 2 MATLAB format 10 5 1 3 Hierarchical data format 10 5 2 Data Visualization and Inspection 13 5 3 Data Processing and Outputs results from Standard runs 16 5 3 1 Covariance Evaluation 16 5 3 2 Merging Procedure 18 EXAMPLE COVIS 3 0 RUNS 20 REFERENCES 23 3 COVIS 3 0 User Manual 1 INTRODUCTION With an increasing availability of satellite time series imagery it has become possible to monitor temporal and spatial variability of coastal and open areas Satellite data provides a great improvement over the more limited spatial sampling offered by ship and in situ such as a buoy Boussole mooring or autonomous vehicles observational platforms Generally merging remote sensing data with in situ measurements has become a standard procedure to increase the quality of satellite derived products Conventionally covariance analysis is applied to oceanographic and meteorological data sets to decompose space and time distributed data into modes ranked by their temporal variance while optimum sampling analysis is applied to find adequate numbers of in situ data to improve satellite quality by reducing their observational error COVIS 3 0 is a Toolbox for implementing such concepts in particular it can be used to Evaluate the covariance from satellite imagery Merge of satellite and in situ measurem
12. The approach is based on an isotropic diffusion operator that preserves frontal structures 2 vyl K WE YO V Fi Vy dt 1 0 where y x t represents the oceanographic measured field from the satellite at a given time and with a known observation error in other words the image Using an optimization process that finds the Gaussian residual field x the covariance of x is assumed to be a function of the relative distance between two points 2 eea e le x This property will allow uncertainties to be computed at different locations without performing real in situ measurements The Covariance in Equation 2 is computed numerically and fitted to a covariance model Gaussian exponential spherical etc To better explain this concept we present a covariance example using an image of Sea Surface Temperature SST with the Advanced Very High Resolution sensor AVHRR onboard NOA remote sensing satellite this image have been provided with COVIS 3 0 COVIS 3 0 SAMPLE_IMAGE GEOTIFF 2009 0606 1144 n18 sst gtiff In particular we present result from the working area with min LON 14 2 max LON 14 65 min LAT 39 2 and max LAT 39 65 To better appreciate the temperature fronts we have modified the colormap min value 21 max value 25 5 decimal degrees That sub area was decomposed using Equation 1 into a spatially varying mean and Gaussian Noise fields as showed in Figure 10 That covariance ev
13. aluation was performed using a default satellite standard deviation of 0 5 that represents the satellite error However if this is known a priori from in situ match up the user can alter 16 COVIS 3 0 User Manual eee COVIS3 0 eee akc 2009 0606 1144 n18 sst gtff 39 std_obs o0 min LON 142 std_sat 05 Latitude deg max LON 14 85 min LAT 39 2 max LAT 39 65 Covarianza 13 9 14 14 1 14 2 14 3 14 4 14 5 14 6 14 7 14 8 14 9 23 23 5 24 24 5 25 25 5 To change the colormap insert here minimax values 21 255 l Display Image Color Display Image Gray Figure 9 COVIS results after the STEP 2 The black box overlays the user selected area size 54 by 53 kilometres Residual Field __ _ Residual s Histogram Figure 10 Mean field and Gaussian Residual field X of the selected area E X is a Gaussian realization with a mean of 0 0047 and a standard deviation of 4 0 0047 and a standard deviation of 0 2047 The covariance was evaluated using Equation 2 and fitted to a covariance model as displayed in Figure 11 where the value at zero spatial lags is obtained by an exponential fitting of the obtained curve In particular in this case the fitting law is given by the following equation 17 COVIS 3 0 User Manual 12 7 C X 0 0145 e 47 3 Figure 1 Fie Edit View Insert Tools Desktop Window Help x DB k Raat 08 50 x10 Covari
14. ance Result cf Covariance 30 r Distance Km Figure 11 Covariance in terms of the spatial lag obtained on the clear SST area of Figure 10 The wave like behaviour starts around 14 kilometres It s important to note that the observed wavelike behavior for distances bigger than 14 kilometers is an artifact of the longitudinally banded structure in the original image that depends of the sensor error This step can be repeated on different areas clicking on Reset Lat Lon while it s also possible to load a new image clicking on the button Reload New Sat Image The waiting time of this computation depends on the size of the selected area if any sub area are identified the covariance will be evaluated on the entire image 5 3 2 Merging procedure Merging remote sensed observations with in situ measurements is a standard procedure to improve the quality of satellite derived products The approach is to study spatial temporal variability in the satellite imagery and to distribute in situ sampling over the image following the covariance criterion Therefore once the covariance C has been obtained from Equation 2 a new field merging in situ and satellite data is retrieved maximizing the following probability distribution P Y g exp o HW o Vos HWk Vk y COW y 4 where y represents the estimated pixel value y is the observation vector H is the obs observation matrix and
15. ance analysis we suggest sizes sub areas no bigger than 60 x 60 pixels For the above explained reasons a requirement for user input requires considerable forethought and effort on the user s part in order to select an appropriate working area and an appropriate size and inputs such as latitude and longitudes corners COVIS 3 0 has been a work in progress since G Pennucci and A Alvarez first started working out the numerical MATLAB algorithms at the core of COVIS 3 0 and it will continue to be so The invariant imbedding MATLAB algorithms at the core of COVIS 3 0 are mature and well debugged after many simulations and tests However many features of the Graphical User Interface GUI are new with Version 3 0 and may evolve quickly as feedback is received from the users 2 SYSTEM REQUIREMENTS COVIS 3 0 s source code is written and compiled entirely in standard MATLAB in order to make the code as portable as possible The GUI was developed with the 7 5 0 342 R2007b MATLAB Version The minimum system requirements for COVIS 3 0 are as follow Operating System Microsoft Windows 2000 2003 XP Vista or MAC OS Processor CPU 1 GHz 32 bit x86 Memory RAM 1 GHz Free Disk space 50 MB MATLAB version 7 or superior MATLAB Toolbox Mapping Toolbox Intel 3 1 COVIS 3 0 GUI is designed for optimum use on monitors with 2014 x 768 pixels Lower resolution monitors cause some of the GUI forms to display as larger than the monitor screen
16. d with APS software COVIS 3 SAMPLE_IMAGEHDF_HDFAMODIS npp 2011326 1122 D L3_Mosaic viirs NEA_npp v02 750m hdf COVIS 3 0 SAMPLE_IMAGE HDF_HDF4AWVIIRS npp 201 1326 1122 D L3_Mosaic viirs NEA_npp v02 750m hdf COVIS 3 0 SAMPLE_IMAGE HDF_HDF4WVIIRS npp 2011326 1122 D L3_Mosaic viirs NEA_npp v02 750m hdf 5 2 Data Visualization and data Inspection Once time a parameter was selected the visualization program displays the satellite image as it is in the original file this means with flags colorbar and color scale as they have been stored in the source file However these parameters can be manually changed using the button in the lower side of the GUI For example in the MERIS HDF file test in covis 3 0 sAMPLE_IMAGE HDF_HDF4MERIS MER_FRS_2PNPDE20090317 hdf the physical parameter available for visualization is the chlorophyll chl as showed in Figure 7 It s possible to note that the chl image is not clear because it has a bad color scale shapes and fronts structures are not well represented due the land flags For this reason the user can decide to explore the image changing the minimum and maximum value of the color scale and or the color map In Figure 8 have been represented the 13 COVIS 3 0 User Manual result changing the maximum values of chl from the land flag high value to the expected and more consistent value of ch 0 8 in color and black amp white version COVIS 3 0 e 46 45 std_obs 001 44
17. ents Generate uncertainty maps and uncertainty indices from the available time series of images and historical information These methodologies can be applied to several optical satellite data sets such as the Moderate Advanced Very High Resolution Radiometer MODIS the Medium Resolution Imaging Spectrometer MERIS the Advanced Very High Resolution Radiometer AVHRR and the Visible Infrared Imager Radiometer VIIRS sensors as well as several optical in situ platforms such as the AErosol RObotic NETwork AERONET and the Marine Optical BuoY MOBY platforms COVIS 3 0 can be used to assess the locations of where in situ data should be collected in order to determine the uncertainty when using these data for their calibration and validation efforts The main goal of COVIS 3 0 is to merge remotely sensed data with the available in situ measurements thus increasing the quality of satellite derived products This User Guide s is independent from any other publication and should be adequate for users who whish to run COVIS 3 0 as a black box to visualize and processing satellite data This User s Guide for assumes that the reader is familiar with the basic terminology and notation of optical oceanography and statistical analysis If this is not the case then the reader should first consult the review manuscript by G Pennucci A Alvarez and C Trees available at NURC or the paper by G Pennucci G Fargion A Alvarez C Trees and
18. g on the button Display Image Gray Once time the image was ingested and visualized represented the user has to define a working area of interest to zoom on a particular area click on the button Zoom in for data processing This step is very important for a proper covariance analysis therefore we have analysed different scenario using Monte Carlo simulations to define the following guidelines that have to be considered by the user for a proper selection of the area of interest 1 Due to computational reasons we suggest sizes no bigger than 60 by 60 kilometres that for MODIS acquisition means about 60 x 60 pixels this because the A optimum calculation requires a number of computation proportional to the size of the area this operation will take some time depending on the speed of your machine s CPU This limitation in size it s also due to the intrinsic definition of the covariance on a satellite image don t make sense to find correlation between pixel that are so far 2 Due to statistical reasons the covariance computation has to be performed on cloud and noise free areas with a coverage lt 10 15 This consideration was the result of several simulations for the evaluation of the covariance field as a function of cloud and noise coverage for more details see Reference 3 The results show that the analysis of the covariance depends on clouds noise size and position In other words we have demonstrated that the in
19. ile is selected from the user the MATLAB file data ingestion folder will appear as in Figure 3 The user has to type the name of the three required input as have been stored in the original file name are case sensitive and click on the button Display Image to visualize and to ingest the image in COVIS 3 0 MATLAB file data ingestion aE Insert variables name Image a grid of size MxN Longitude vector size 1 x N Latitude vector size 1 x M Display Image Quit Figure 3 MATLAB mat data ingestion procedure 5 1 3 Hierarchical Data Format HDF HDF4 and HDF5 hdf hdf4 hdf5 or h5 Hierarchical Data Format is the name of a set of file formats and libraries designed to store and organize large amounts of data such as remote sensing satellite products In 1993 NASA chose the Hierarchical Data Format Version 4 HDF4 to be the official format for all data products derived by the Earth Observing System EOS In 2008 due to some limitations the HDF4 format was improved in terms of data compression and organization and a new HDF format was introduced HDF5 Because HDF4 and HDF5 are completely independent formats the calling interfaces of COVIS 3 0 were different Unfortunately this is not enough to ensure consistent data 10 COVIS 3 0 User Manual ingestion because the storage formats of HDF HDF4 HDFS5 is different and depends on several the processing Level of the stored products and from the proces
20. nce NATO vessel during a cruise that was performed at the same time of the AVHRR acquisition STEP 4 To explore the image and to define the area of interest you have to click on the Zoom in and the right mouse button can be used to select the region of interest The Min Max values of the color scale are displayed automatically but they can be changed inserting the desired values on the lower part of the window as in Figure 13 STEP 5 Once time the sub area was identified for the processing covariance evaluation four corners have to be inserted In this case we focus the attention on longitude from 9 2 to 9 7 min and max LON and latitude from 43 4 to 44 1 min and max LAT as showed in Figure 14 NOTE THAT once time the black box area was defined and displayed in the GUI it s not possible to change the colormap values STEP 6 To evaluate the covariance click on the Covariance button The calculation time depends on the size of the area selected and the covariance results will be displayed in a GUI figure Figure 15 that displays the covariance length for the area STEP 7 Once the covariance have been evaluated it is possible to merge the previously loaded in situ data with the original satellite image Figure 16 displays the final results of merging 20 COVIS 3 0 User Manual Load Satellite Image Load in situ data std_obs 0 01 std_sat 05 Latitude deg Covarianza Display Merged Image
21. nput images to generate covariance map Such information is fundamental to merging satellite image with in situ and to perform uncertainty analysis of the area of interest COVIS 3 0 can be run with several types of in situ data to merge remote sensing acquisitions with the available in situ measurements As written before COVIS 3 0 is a work in progress tool and additional MATLAB algorithms are ready to be integrated in the core of COVIS 3 0 In particular we are working on the possibility to run COVIS 3 0 on time series of satellite imagery tridimensional array to perform uncertainty analysis This function will allow the user to automatically produce uncertainty maps that can be used to characterize the background environments in a particular month for more details see References 1 and 2 4 2 Inputs In order to run COVIS 3 0 to evaluate the covariance distribution of a particular image during a particular environmental condition the user supplies the core mode with the following information or via direct input or via user written or via user supplied data files A satellite image To directly ingest the image of interest the file format must be Geotiff HDF4 or HDF5 Otherwise the user can provide a MATALB mat file format in which was saved as a grid or vector the longitude and latitude X and Y 6 COVIS 3 0 User Manual respectively information and a grid with the same X and Y size for the image pixels physical
22. nt the attention on one important common property Because COVIS 3 0 was written to process georeferenced data for a proper covariance merging and match up analysis the satellite and in situ file used have to contain GIS standard products In other words the working file has to contain at least one projected image and the relative latitude and longitude information Lookin E ovs 5 same e_Imace My Recent To select the desired My Zomputer file eg lt MyNawak Fie rams laces Fies of ype Geotiff aeoti Files I aiff aif i Load Image eles Look in E SAMPLE_IMAGE gt e 0ce 3 My Recent Docunents gt 3 Desktop u3 My Documents Melee MyConpute gt CI 3 Floppy lA To select the lt Local Disk C 5 Soa Path of the ree E Documents and Settings desired file 3 D 2ennucci Desktea a My Documents SAMPLE_IMAGE 2 DVD RAM Drve fE 2 My Network Places cous My Computer tools ee E IIRS reader My Network Filenaine Frar z Oen Places Files oftype HIF4 HDF5 Files nd z Cancel Figure 2 Remote sensing data ingestion from using COVIS 3 0 3 0 5 1 1 TIF and GeoTIFF format tif gtiff or gtif This type of format is used by over different remote sensing companies and organizations as an interchange format for georeferenced raster imagery The potential additional information is stored in a structure and it includes map projection
23. o test this GUI procedure we have provided two examples of HDF files that contain MERIS as showed in Figure 5 and MODIS satellite data These files tests have been provided with the COVIS 3 0 MATLAB program under following directories COVIS 3 0 SAMPLE_IMAGE HDF_HDF4 MERIS MER_FRS_2PNPDE20090317 hdf COVIS 3 0 SAMPLE_IMAGE HDF_HDF4 MODIS MODPM2009076125154 hdf b Standard HDF HDF4 Level 3 data file generated by APS This function was specifically written to process HDF HDF4 files that have been generated by the Automated Processing System APS APS is a collection of programs designed to generate co registered image databases of geophysical parameters derived from remotely sensed images One of the main characteristic of that system is the automation that is the technique of making a system operates without human effort or decision This procedure allows the generation of standard HDF structure same names for geographic data same map projection efc that can be read using automatic procedures For this reason we have decided to develop a specific GUI that can be use to visualize and ingest APS Level 3 products in COVIS 3 0 As displayed in the following Figure this GUI is faster and automatic to respect the GUI developed for the generic HDF file Indeed in that case the user has to select only the physical data of interest that has to be a two dimensional grid of data while the geographic parameters are automatically ingested in COVIS 3
24. sing tool used for their generation such as APS ADR SeaDAS etc For this reason COVIS 3 0 was developed to ingest several of the more used HDF HDF4 HDFS5 as listed below a Generic HDF HDF4 data file when the processing tool is unknown b Standard HDF HDF4 Level 3 for files generated with APS c HDF5 VIIRS SDR for files generated with ADL 3 1 d HDF5 VIIRS EDR for files produced with IDPS a Generic HDF HDF4 data file This case represents the more general situation in which the user doesn t know the HDF source provider This is a general interface that requires expert users because they have to manually select the appropriate variables for data ingestion and processing In particular when a generic HDF hdf file is selected the HDF data selection GUI will appear as showed in Figure 4 This GUI requires three inputs that have to manually typed by the user choosing the variables from those available listed in the left side of the screen Note that name are case sensitive _ HDF data selection 4 x List of data in the HOF file AAOT al Insert variables name c CP_Latitudes CP_Lines CP_Longitudes CP_Pixels FILENAME IMAGE LAT LAT_box Display Image LON LON_box La_412 La_443 Reload new layer La_488 Figure 4 Generic HDF HDF4 hdf GUI for data ingestion procedure In particular the name of the physical parameter of interest has to be written in the Image box for example La_488
25. t parts COVIS 3 0 oeo Data Ingestion Load Satelite Image Load in situ data std_obs 0 01 H side GS a ae ae in insert min LON deg stdsat 05 aft 3 insert max LON deg insert min LATIdeg T kJ 3 E D insert max LAT deg 100 150 0 0 50 200 Lente deg Data Visualization To change the colormap insert here minimax values min val maxval Display image Color Display image Gray LD Figure 1 COVIS 3 0 MATLAB GUI for covariance estimation and merging 5 1 Data Ingestion COVIS 3 0 requires two files as input for covariance and merging computation Remote sensing satellite data Clicking on the button Load Satellite Image a folder will appear the user has to select the desired file format and to find the file of interest navigating from the Look in search bar as displayed in Figure 2 Allowed file format GeoTIFF TIF HDF4 HDF5 and MATLAB file format In situ data if available Clicking on the button Load in situ data a folder will appear as above the user has to select the desired file format and define the path of the file of interest in the Look in search bar Allowed format TEXT EXCEL and MATLAB file format NOTE in situ data are needed just for merging and data match up but they are not required for covariance analysis COVIS 3 0 User Manual Before to describe in detail each procedure for the different satellite format it s important to poi
26. tellite Covariance Based Method to evaluate the variability in uncertainty maps generated from satellite ocean color imagery NURC report 2 G Fargion G Pennucci A Alvarez C Trees R Arnone A methodology for calibration of hyperspectral and multispectral satellite data in coastal areas SPIE paper number 8372 19 3 G Pennucci A Alvarez and C Trees Study of covariance as a function of cloud cover NURC report under progress December 2010 23 COVIS 3 0 User Manual
27. troduction of not valid pixels causes a loss of information and decrease the correlation length Once time the working area was identified it has to be defined inserting four latitude and longitude corners decimal degrees in the right side of the GUI After the 4 corners introduction it s possible to verify the selection clicking on the button display sub area a black box will appear The selected area can be changed clicking on the button RESET LAT LON and inserting the new corners 15 COVIS 3 0 User Manual 5 3 Data processing and outputs results from standard runs As described in the introduction of this User Manual the core of our study is the evaluation of the covariance matrix of satellite images to describe the spatial variability of the area of interest This information can be also used to increase the quality of the satellite products using in situ observational resources if available for merging with the image under study In this paragraph we briefly introduce methods and algorithms that we have developed more details can be found in references 2 and 3 to explain in details the procedures to implement them with COVIS 3 0 5 3 1 Covariance evaluation One of goals of the COVIS 3 0 is to evaluate the covariance of satellite images to describe their spatial variability To accomplish this we have developed a method to decompose the observed fields into spatially varying mean and Gaussian residual fields
28. values In situ data if available The user can provide k in situ measurements that have to be stored in a MATALB mat as a grid of k rows by 3 columns with longitude latitude and in situ physical value respectively 4 3 Outputs Output from COVIS 3 0 consists of both print out a MATLAB figure to view to save and or to print final results and digital data stored in MATLAB format mat The default printout gives a moderate amount of information showing the standard the quantities of interest such as covariance map merged and difference images and uncertainty maps Figures are useful for taking a quick look of results or for cutting and pasting a particular part of the output into another document If the user is interested in processing the outputs or in a more detailed analysis the MATLAB variables stored in the MATLAB workspace can be used 7 COVIS 3 0 User Manual 5 GRAPHICAL USER INTERFACE ORGANIZATION In this Chapter we describe in detail the processing context of COVIS 3 0 that defines all the parameters necessary to process satellite and in situ data for covariance computation and data merging The processing context is made of up three parts Data ingestion Data visualization data inspection sub area definition for covariance computation Data processing and outputs results Figure 1 displays COVIS 3 0 GUI with its main buttons that have been regrouped following the above described contex
29. while the Lon and Lat empty fields require the name of the variables that contains longitudes and latitudes information in decimal degrees It s important to note that the size of the longitude and latitude information has to be consistent with the size of the selected image In the following table are resumed the size allowed on the basis of the HDF files processed until now IMAGE LONGITUDE SIZE LATITUDE SIZE SIZE CASE 1 mxn ixj withij lt m n ixj withi j lt m n CASE 2 mxn Ixn lxm 11 COVIS 3 0 User Manual Once time these empty fields were compiled the user can store and visualized them in COVIS clicking on the button Display Image as showed in Figure 5 After this procedure the selected parameters are ready to be processed for covariance analysis and or in situ match up as explained in detail in Chapter 6 z improving_sat_info HDF data selection Load Satelite Image aqua 2009244 0901 120149 D L3 mod List of deta in the HOF fie H AAOT c CP_Latitudes CP_Lines Load in stu data H CP_Longtudes Lon CP_Longitudes Lat CP_Letitudes stdobs 00 std_sat 05 a a Py 3 Gi 3 Display Merged image Reset LAT LON Reload New Set Image To change the colormap insert here minimax values 0 003 Display Image Color l Display Image Gray Figure 5 HDF data viewer of COVIS 3 0 with a user selection La_488 and data visualization T
Download Pdf Manuals
Related Search