This section contains all significant changes for each PolSARpro release.
- Calibration assessment procedure: a new functionality which checks the data and shows if there exists azimuth squint
- UAVSAR quad-pol data reader
- A complete link with MapReady (ASF) which can be used to geocode data. This performs ALOS-PALSAR (quad-pol and dual-pol), RS2 (quad-pol and dual-pol) and TSX (dual-pol). The ouput data format is T3 or C2.
- A collection of new functionalities for multi-temporal data processing
- R.O.I extraction
- Mask functionality: define a mask and apply it on the data
- Coarse-registration for dual-pol data
- Orientation estimation (JS Lee and Y Yamagushi algorithms)
- RVOG Polar (Shane - EUSAR2010)
- Texture analysis
- Texture statistics
- Surface inversion: Oh2004 and X-Bragg 2008
- This new version of PolSARpro is able to detect automatically what kind of polarimetric data format the user wants to process
- Bugs fixing of version 4.1.5
Changes from Version 4.0 Beta 1 to Version 4.0 Beta 1.3
Release 4.0 Beta 1.3 fixes the following known problems of Version 4.0 Beta 1 version:
- Quegan Calibration
- 4 pixels shift Radarsat2 GeoTiff Data Format
- Radiometric Calibration in comparison with ASF data product
- Calibration DN factor (amplitude / Intensity) – ASAR APS APP APG and ALOS PALSAR data products
- [T4] data processing flow
- Radarsat2 / TerraSAR-X: file structure with long path name or path name including “space” character
- Radarsat2 Dual Pol – EOPI Generic PolSARpro button
- Radarsat2 Data Format (LE – BE)
- Oh and Dubois surface inversion procedures (incidence angle)
- Hierarchical segmentation procedure
Changes from Version 3.0 Beta 3.34 to Version 4.0 Beta 1
1) New Design of the PolSARpro Software
The new version of the PolSARpro Software (Version 4.0 Beta 1) has been delivered during POLINSAR 09 Workshop (January 2009)
The new version of the PolSARpro Software (Version 4.0 Beta 1) proposes a complete new design of the main interface. In the previous version, the PolSARpro Software was composed of a full collection of independent widgets associated to a given functionality. In this new version (Version 4.0 Beta 1), all these widgets are now seen as transient windows by the Window Manager and so, are all embedded in only one main window. Minimizing or maximizing this Master Window will minimize / maximize or iconify all opened transient widgets.
2) New PolSARpro Features:
Interface with External Software
The PolSARpro v4.0 software can now offer the possibility to interface with the following external softwares:
- G.I.M.P (Gnu Image Manipulation Program): This open source software, under GNU License, can handle and display several BMP images at the same time, thus allowing a better comparison between them.
- Google Earth: Different functionalities are also provide that offer to the user the possibility to visualize on Google Earth the SAR Image Track (Read / Create / Open a GEARTH_POLY file), or overlay and display on Google Earth the PolSARpro output BMP images (Create / Open a KML file).
- S.R.T.M (Shuttle Radar Topographic Mission): This functionality offers to the user to download from two given ftp sites, the SRTM GeoTIFF file and display and overlay on Google Earth the extracted DEM image corresponding to the SAR image under process.
- N.E.S.T (New ESA Software Tool) : A new link has been implemented in the PolSARpro v4.0 Software to launch directly the ESA-NEST Software, thus proposing to the user all the G.I.S functionalities offered by the NEST Software.
Import
The PolSARpro v4.0 software can now offer the possibility to handle and convert polarimetric data from the ALOS / PALSAR S paceborne Sensor, according to the ERSDAC ( Earth Remote Sensing Data Analysis Center ) – Vexcel Format .
PolSARpro – Simulators
From the PolSARproSim simulator developed by Dr Mark L. Williams and providing simulated Pol-InSAR test data on forests, two new simulators on real environments are proposed:
- PolSARproSIMgr : for Pol-InSAR ground surface simulation
- PolSARproSIMgv : for Pol-InSAR ground + small vegetation simulation
Single Data set - Data Process
Different additional functionalities are proposed in the release Version 4.0 Beta 1.3:
- Box-Car + Edge detectors : This speckle filter uses the outputs of the edge detector functionality, thus preserving linear features or structures during the filtering process.
- C Lopez and Lee-Sigma speckle filters : Two new speckle filters, recently published by C. Lopez and J.S. Lee, are implemented.
- New Polarimetric Parameters :
- Conformity Coefficient (M.L. Truong Loi)
- Differential Reflectivity (ZDR)
- Polarimetric Reflectivity Ratio
- Scattering Predominance (J. Praks & E. Colin)
- Degree of Purity (J. Praks & E. Colin)
- Scattering Diversity (J. Praks & E. Colin)
- Depolarisation Index (J. Praks & E. Colin)
- Entropy approximation (J. Praks & E. Colin)
- Alpha approximation (J. Praks & E. Colin)
- Scattering Mechanism Entropy (A. Freeman)
- Kozlov Anisotropy coefficient
- Faraday Rotation Estimation
- M. Neumann 2-components decomposition
- R. Touzi - TSVM decomposition
- J.J. VanZyl 3-model based components decomposition (published in 2008)
- Unsupervised fuzzy neural segmentation ( S.E. Park )
- Algorithms for topography extraction from PolSAR data.
- Surface parameter data inversion: Oh 2004 Model.
- Rewrite ENVI Headers for PolSARpro Classification bin files incorporing the color tables used to generate the BMP files. ( Courtesy of Dr. Joseph Buckley from Royal Military College of Canada – Kingston – Canada )
Multi Data sets – Dual POLinSAR – Data Process
- Pol-InSAR Coherence Optimisation Methods: New approaches, proposed by M. Lavalle are implemnted: Singular Value Decomposition (SVD), Phase Diversity (Tabb), Numerical Radius (E. Colin), Min/Max of Phase / Magnitude over Coherence Boundary and Max Phase / Max Difference over Coherence Boundary.
- Baseline Calculator Tool (M. Lavalle):
- Estimation of the horizontal, vertical, parallel and perpendicular baselines from the state vectors
- Coarse Co-Registration of the [S2] SLC PolSAR images
- Generation and estimation of the Flat Earth map
- Generation and estimation of the Vertical Wavenumber (kz)
Additional Modules Development
An automatic procedure, called configuration has been implemented, that checks during the installation what is the PDF Reader software installed in the machine of the user, and creates the associated link.
This automatic procedure also checks and creates a link between PolSARpro v4.0 Software and Google Earth and GIMP functionalities if installed on the machine of the user.
Changes from Version 3.0 Beta 3 to Version 3.0 Beta 3.1
1) New PolSARpro Features:
Multi Data sets – Dual POLinSAR – Data Process
-
Polarization Coherence Tomography (PCT): this new and original approach, proposed by Prof. S.R. Cloude, concerns a new radar imaging technique which employs variation of the interferometric coherence with polarization to reconstruct a vertical profile function in penetrable volume scattering.
S. R. Cloude, “Polarization Coherence Tomography”, Proceedings of 6 th European SAR Conference , EUSAR 06, Dresden , May 2006
S. R. Cloude, “Polarization Coherence Tomography”, Radio Science , 41, RS4017, September 2006
S. R. Cloude, “Dual Baseline Coherence Tomography”, IEEE Geoscience and Remote Sensing Letters , Vol4, No. 1, January 2007, pp 127-131
Changes from Version 3.0 Beta 2 to Version 3.0 Beta 3
1) New PolSARpro Features:
The PolSARpro v3.0 software can now offer the possibility to handle and convert polarimetric data from the TerraSAR-X Spaceborne Sensor.
The TerraSAR-X toolbox is able to import data from the following Dual-Pol modes:
- SSC : Single Look Slant Range Complex
- EEC : Enhanced Ellipsoid Corrected
- GEC : Geocoded Ellipsoid Corrected
- MGD : Multi-Look Ground Range
The PolSARpro v3.0 software package proposes a specific interface controlled through a graphical user interface (GUI) that can help and guide an EO investigator to handle and process TerraSAR-X Dual-Polarimetric data sets along a ‘recommended' processing chain using only the following functionalities: Quicklook, Extract Sub Area, Data Conversion, Header Analysis and Data Processing. The proposed Data Processing functions are in the form of one (or more) batch processing, covering:
- Convert Complex Dual-Pol data to 2x2 covariance matrix ( C2 )
- Stokes parameters analysis
- Speckle Filter (BoxCar, Gaussian, JS Lee refined filter, IDAN, PWF)
- Edge detector
- Data clustering
- H / A / a Decomposition and analysis
- Supervised Wishart Classification
In the PolSARpro v3.0 Software, these functionalities have also been extended to handle TerraSAR-X data, thus taking into account coherent dual-polarimetric data ( Spp ) or incoherent dual-polarimetric data ( [C2] ).
Single Data set - Data Process
Different additional functionalities are proposed in the release version Beta 3:
-
Edge detectors : Functionality that can be used to develop a CFAR linear fueature extraction, like roads or urban structures. Four different edge detector procedures are proposed: Canny, Black, Marr and Rothwell.
-
Data Clustering : This functionality based on a watershed procedure approach is helpful for pure class selection and improves classification results .
-
Polarimetric Entropy according Shannon theory : This function is a courtesy of Jerome Morio and Philippe Refregier from University of Aix – Marseille III (Institut Fresnel - Marseille).
-
Coefficient of Variation
-
Lueneburg Anisotropy coefficient
-
Generalized Freeman – Yamaguchi 3-components decomposition
-
Yamaguchi 4-components decomposition (published in 2007)
-
Freeman 2-components decomposition (published in 2006)
-
Unsupervised segmentation based on the generalized Freeman – Yamaguchi 3-components decomposition
-
A new 24bits RGB BMP representation of the dual-polarimetric data based on the Stokes parameters
-
A new 24bits HSV BMP representation of the dual-polarimetric data based on the Stokes parameters
PolSARpro Viewer
A new functionality, based on the ERDAS View Finder , has been added to the PolSARpro Basic Viewer that offers in the same graphic widget, the Display of the BMP image, the Lens functionality and the Zoom functionality.
Statistics
The Statistics Work Package, already implemented in the PolSARpro v3.0 Software, has been extended to be available to any Dual-Pol input data format : [Spp] and [C2] .
Additional Modules Development
An automatic procedure, called configuration has been implemented, that checks during the installation what is the PDF Reader software installed in the machine of the user, and creates the associated link. This automatic procedure also checks and creates a link between PolSARpro v3.0 Software and Google Earth functionality if installed on the machine of the user.
2) Bug Fixes:
The Beta 2 release had the following known problems that have been solved in the Beta 3 Version:
Single Data sets – Data Process
Import
PolSARpro – Simulator
PolSARproSim , developed by Dr Mark L. Williams , is an educational tool providing simulated POLinSAR test data of sufficient fidelity to be used within the tutorial package of PolSARpro v3.0 .
Some numerical problems have been solved and the new version is provided.
Changes from Version 3.0 Beta 1 to Version 3.0 Beta 2
1) New PolSARpro Features:
Import
The ALOS-PALSAR interface has been extended to handle and convert Dual-Pol polarimetric data. The ALOS-PALSAR is now able to import data level 1.1 and data level 1.5 Quad-Pol mode and Dual-Pol mode. The ALOS-PALSAR interface automatically extracts now the Quicklook image included in the TRAILER FILE. The corresponding BMP file is saved under the name:
low_resol_quicklook.bmp
Single Data set - Data Process
Different additional functionalities are proposed:
- IDAN Speckle Filter: This function corresponds to the Intensity Driven Adaptive Neighbourhood speckle filter and is made available to any input data format and polarimetic representation: [S],[T3], [T4], [C3] and [C4].
This function is a courtesy of Gabriel Vasile and Emmanuel Trouve from University of Savoie Annecy
- Compact Polarimetry: Recent interest in dual-pol SAR system has lead to a novel approach to dual-pol SAR, the so-called Compact Polarimetric Mode. (see recent publications by J.C Souyris et al., T. Ainsworth et al., N. Stacy et al., K. Raney). This new technique shows the promise of being able to reduce the complexity, cost, mass and data rate of a SAR system while maintening capabilities of a fully polarimetric system. This function offers the possibility to perform a Compact Polarimetric Mode from the (2x2) complex Sinclair monostatic [S2] or the (3x3) complex Coherency [T3] or Covariance matrix [C3] polarimetric data sets. Such compact dual-pol polarimetric SAR system collects scattering vectors by transmitting a given polarized signal and the coherently receiving any pair of orthogonal polarization basis. These corresponding scattering vectors are thus constructed by the way of Special Unitary operators from SU(2) or SU(3) and the results are saved in a (2x2) Covariance matrix ([C2]) raw binary data format. This function also allows construction of pseudo quad-pol information from the generated dual polarization raw binary files, and save the results either in a (3x3) Covariance matrix ([C3]) or in a (3x3) Coherency matrix ([T3]) format raw binary data format
Multi Data sets – Dual POLinSAR – Data Process
A complete reorganization of the different POLinSAR functionalities that can be applied both to 2x2 complex Sinclair matrix [S2] or 6x6 Coherency matrix [T6] raw binary data, offering now two new dedicated widgets:
- Coherence Estimation
- Height Estimation – Inversion Procedures
- Coherence Estimation: Creates complex coherences binary files at any possible polarization states, including the Lexicographic (Linear), Circular, Pauli basis and optimal polarization states derived from Cloude and Papathanassiou coherence maximization procedure.
- Height Estimation – Inversion Procedures: All the proposed Height Inversion Procedures are now grouped together in the same functionality.
- Phase Center Height
- DEM Differencing Algorithm
- Coherence Amplitude Inversion
- Ground Phase Estimation
- Phase and Coherence Estimation - RVOG Algorithm
These different inversion algorithms are all based on both the use of the two selected polarisation channels (now directly selected from the corresponding widget) and the scale factor or vertical wavenumber kz.
Tutorial On Radar Polarimetry and Radar Polarimetry Interferometry
The POLinSAR Training Course, illustrated with the POLinSAR tutorial that can be used as a low level teaching aid for polarimetric SAR interferometry processing, has been updated according to the new corresponding POLinSAR widgets that the Tool offers.
Open Source Development
All the development languages and compilers used are open source and completely free languages. Any C routine is completely independent of the system environment and can be simply and directly compiled with the included graphic and/or mathematic linked library. Users have access, directly from the PolSARpro Software main interface, to a complete description of all C Routines provided within the software. The list of all the C routine documentation file, that proposes a complete description of the source code, has been fully completed and updated.
Softawe Portability
The PolSARpro v3.0 software was already running on all of the following platforms: Windows 98+, Windows 2000, Windows NT 4.0, Windows XP, Linux Red Hat.
The PolSARpro v3.0 – Beta 2 software version has been upgraded to be able to run on an UNIX - Solaris 2 platform.
2) Bug Fixes:
The Beta 2 release had the following known problems that have been solved in the Beta 1 Version:
Multi Data sets – Dual POLinSAR – Data Process
PolSARpro – Simulator
PolSARproSim, developed by Dr Mark L. Williams is an educational tool providing simulated POLinSAR test data of sufficient fidelity to be used within the tutorial package of PolSARpro v3.0.
Some numerical problems have been solved and the new obtained results are now consistent with the POLinSAR tutorial examples.
Changes from Version 2.0 Beta 8 to Version 3.0
1) New PolSARpro Features:
Import
The PolSARpro v3.0 software can now offer the possibility to handle and convert polarimetric data from the Spaceborne Sensors:
- ALOS – PALSAR: data level 1.1 and data level 1.5
- RadarSAT-2 Dual-Pol mode and Quad-Pol mode.
Earth-Observation Scientific Investigator Toolbox
The PolSARpro software package proposes different specific interfaces dedicated to different Spaceborne Sensors (ALOS-PALSAR, RADARSAT2, TerraSar, SIRc) or Airborne Sensors (AIRSAR, Convair, EMISAR, ESAR, PISAR, RAMSES) data processing. All these different specific interfaces are controlled through the same graphical user interface (GUI) in such a way that it can help and guide an EO investigator to process Single Polarimetric Data Set along a ‘recommended’ processing chain using only the following functionalities: Quicklook, Extract Sub Area, Data Conversion, Header Analysis and Data Processing. The proposed Data Processing functions are in the form of one (or more) batch processing, covering:
- Convert Complex data to 3x3 coherency matrix (T3)
- Speckle Filter (BoxCar, JS Lee refined filter)
- H / A / a Decomposition and analysis
- Unsupervised Wishart Classification
- Supervised Wishart and/or Gaussian Classification
In the PolSARpro v3.0 Software, these functionalities have been extended to handle ALOS-PALSAR, RadarSat-2 and future TerraSAR-X data, thus taking into account coherent dual-polarimetric data (Spp) or incoherent dual-polarimetric data ([C2]).
Single Data set - Data Process
Different additional functionalities are proposed:
- Gaussian Speckle Filter: This function is equivalent to a Box Car speckle filter but uses a Gaussian kernel during the filtering.
- Stokes Parameters: This function creates binary files corresponding to the different Stokes parameters constructed both column of the corresponding (2x2) complex Sinclair [s2] raw binary data.
- A complete reorganization of the H/A/a decomposition functionality, offering now three dedicated widgets:
- Decomposition Parameters
- Eigenvector Set Parameters
- Eigenvalue Set Parameters
- Polarimetric Segmentation – Rule-Based Hierarchical Classification: Different approaches have been used to classify SAR data, e.g. the maximum likelihood approach using extracted polarimetric features, the maximum likelihood approach using the Wishart distribution, or the unsupervised approach using polarimetric decomposition. A very important class of algorithms is the knowledge-based approaches (see H. Skriver publications). Here, generic characteristics of different cover types are derived by combining physical reasoning with the available empirical evidence. These are then used to define classification rules. Because of their emphasis on the physical content of the SAR data they attempt to generate robust, widely applicable methods, which are nonetheless capable of taking local conditions into account. The hierarchical decision trees method uses a sequence of decisions to progressively refine the labelling given to a given pixel or field. Broad classes are distinguished first, with each subsequent step creating new subclasses that are more and more specific. Typically, such approaches are based on some physical understanding of the scattering processes giving rise to the polarimetric signatures, and using this to isolate the most relevant features for identifying different cover types. A specific GUI interface has been developed to graphically design the tree architecture.
- Data Analysis: To complement the statistics functionality already implemented, two new functions are proposed: Data Histograms, to plot histogram of any kind of data representation (modulus, phase, real part, imag part, dB etc …) from a selected area, and Data Profiles to get data value and plot data transects from a selected area. Specific graphical widgets have been developed for these two new functionalities.
Multi Data sets – Dual POLinSAR – Data Process
Different new additional functionalities are proposed, and can be applied both to 2x2 complex Sinclair matrix [S2] or 6x6 Coherency matrix [T6] raw binary data:
- Gaussian Speckle Filter: This function is equivalent to a Box Car speckle filter but uses a Gaussian kernel during the filtering.
- Interferogram generation: Creates interferogram files at any possible polarization states, including the Lexicographic (Linear), Circular and Pauli basis.
- Coherence Estimation: Creates complex coherences binary files at any possible polarization states, including the Lexicographic (Linear), Circular, Pauli basis and optimal polarization states derived from Cloude and Papathanassiou coherence maximization procedure.
- Coherence Analysis – Complex Plane: This Application is used to display both phase and amplitude of the complex coherence inside a unit circle in the complex plane. This is a pixel-by-pixel view and it is an important tool in POLinSAR as it allows viewing the variation of both phase and coherence with polarisation. The selection of points of interest and the results displayed is achieved by the way of an interactive Graphical User Interface
- Coherence Analysis – Coherence Region: This Application is used to display the boundary of the coherence region (both phase and amplitude) inside the unit circle in the complex plane, for all polarisation vectors. (see POLinSAR Training Course) The selection of points of interest and the results displayed is achieved by the way of an interactive Graphical User Interface.
- Height Estimation – Phase center height: This functionality is used to estimate height from the interferogram phase, using the scale factor or vertical wavenumber kz. Users have the opportunity to generate Phase Center Height estimation files at any possible polarization states, including the Lexicographic (Linear), Circular, Pauli basis and optimal polarization states derived from Cloude and Papathanassiou coherence maximization procedure.
- Height Estimation – Coherence Estimation: This functionality is used to choose a priori two polarisation channels that will be used in different height inversion procedures. The first one corresponds to the phase centre closed to the canopy, the second one to the ground.
- Height Estimation – Inversion Procedures: Three proposed Height Inversion Procedures are grouped together in this functionality.
- DEM Differencing Algorithm
- Coherence Amplitude Inversion
- Ground Phase Estimation
- Phase and Coherence Estimation - RVOG Algorithm
These different inversion algorithms are all based on both the use of the two selected polarisation channels (gamma_high.bin and gamma_low.bin) and the scale factor or vertical wavenumber kz.
- POLSAR segmentation: This function is used to apply sequentially different polarimetric data processes. The different steps of this batch process are:
- H / A / Alpha Decomposition and analysis
- Unsupervised Wishart - H / A / Alpha Classification
- Basic Scattering Mechanisms identification
This basic processing approach provides a first qualitative analysis of the fully polarimetric data set processed.
- POLinSAR segmentation: This program creates binary and bitmap image files resulting from the POLinSAR unsupervised classification of polarimetric interferometric data using Wishart Maximum Likelihood statistics.
Further information about this classification algorithm can be found in :
L. Ferro-Famil, E. Pottier, J.S. Lee: “unsupervised classification of natural scenes from polarimetric interferometric SAR data”, in Frontiers of Remote Sensing information processing, C.H. Chen, Singapore, World Scientific Publishing, 2003, pp 105 – 137.
- Data Analysis: Two new functions are proposed to make a POLinSAR data analysis: Data Histograms, to plot an histogram of any kind of data representation (modulus, phase, real part, imag part, dB etc …) from a selected area, and Data Profiles to get data value and plot data transects from a selected area. Specific graphical widgets have been developed for these two new functionalities.
Tutorial On Radar Polarimetry and Radar Polarimetry Interferometry
PolSARpro v3.0 Software proposes a Tutorial to provide a substantial and balanced introduction to the basic theory, scattering concepts, systems and advanced concepts and applications typical to Radar Polarimetry and Interferometry.
The POLinSAR Training Course is now completed with a review of some important aspects of POLinSAR to learn how to generate multiple polarization interferograms and how to use them with simple inversion models to estimate vegetation height. The POLinSAR Training Course can be launched from the TUTORIAL button in the PolSARpro v3.0 Software, then follow Tutorial / Part II – Polarimetric SAR Interferometry.
The POLinSAR Training Course is illustrated with a new POLinSAR tutorial that can be used as a low level teaching aid for polarimetric SAR interferometry processing and is illustrated with application examples showing the full range of functions that the Tool offers.
For this, a specific Do it yourself has been created. This new functionality can be launched from the TUTORIAL button in the PolSARpro v3.0 Software main interface, then follow Tutorial / Part VI – Do It Yourself / POLinSAR Training Course.
In order to improve the general tutorial on Radar Polarimetry, three specific Lectures Notes have been provided, that can be accessed from the TUTORIAL button in the PolSARpro v3.0 Software main interface, then follow Lectures Notes.
PolSARpro – Simulator
The POLinSAR training course is illustrated using a simulated L-Band SAR image of a woody hedge above rough ground. The data was produced using a high-fidelity SAR coherent scattering and imaging code developed by Dr Mark L. Williams. This code has been used extensively with detailed forest models to demonstrate the feasibility of many of the techniques that PolSARPro v3.0 software seeks to illustrate and teach. Simulated data is extremely useful for tutorial purposes as it represents a simplified system that is readily understood, and has well defined properties that make the POLinSAR inversion reasonably straightforward.
PolSARproSim is an educational tool providing simulated test data of sufficient fidelity to be used within the tutorial package of PolSARpro v3.0. The concept is that the user is provided a simple interface that permits the ready description of a realistic forest model and a flexible description of SAR imaging parameters. Having designed the forest and set the imaging scenario the simulation could be called from within PolSARpro v2.0 software, and in a reasonable space of time, the user has available simulated polarimetric, interferometric SAR data for the scene of their choice.
The PolSARpro v3.0 software user could then proceed to follow the POLinSAR tutorial with the data appropriate to the scenario of particular interest to them, and be able to determine that the techniques work well in a realistic situation for the sensor of their choice.
The graphical user interface (GUI) used to assign parameter values is implemented within PolSARpro v3.0 software. From pre-calculated files provided with the simulator, like: a DEM file, a voxel descriptor file, a map of stem locations and an under storey description file, the SAR simulation proper will generate a stem/primary architecture file then calculate the polarimetric, interferometric SAR images.
The technology of the coherent SAR simulation developed by Mark L Williams makes this facility a very real possibility.
2) Bug Fixes:
No specific bug have been raised from the PolSARpro v2.0 Beta 8 version
Earthnet
Print version