Scattering type parameter extraction and a novel clustering scheme for full (FP) and compact (CP) polarimetric SAR data

General information

This code performs an unsupervised clustering using the parameters , for FP and , for CP data. and are the target charactization parameters for FP and CP data, given as,

Here, , and are diagonal elements of T3 matrix. SC and OC are defined as,

and,

;

and are Stokes elements of CP SAR data. and are 3D and 2D Barakat degree of polarizations.

Clustering illustration

The shaded areas are non-feasible regions. Please follow this article for more details: https://doi.org/10.1016/j.isprsjprs.2020.09.010.

Up and running

This is a MATLAB based code. To run the code the coherency matrix elements (T3) for FP and covariance matrix elements (C2) for CP are required.

If and are already in the parent folder then you may use the 'unsupervised_clustering_FP.py' to compute the clustered image. If , , are already in the parent folder then you may use the 'unsupervised_clustering_CP.py' to compute the clustered image.

N. B.

• The T3 and C2 matrix elements should be exported in PolSARpro format and the T3 or C2 directory should contain a "config.txt" file as generated by PolSARpro
• Please specify the window size for processing

In addition to this for CP:

• Please specify the Chi value for processing: for right circular transmit it is -45 and for left circular transmit it is +45

• Please specify the Psi for processing: for CP it is 0

• Please use FP_target_characterization_clustering.m or, CP_target_characterization_clustering.m to perform unsupervised clustering of FP or, CP data, respectively.

References

• S. Dey, A. Bhattacharya, D. Ratha, D. Mandal and A. C. Frery, "Target Characterization and Scattering Power Decomposition for Full and Compact Polarimetric SAR Data," in IEEE Transactions on Geoscience and Remote Sensing, doi: https://doi.org/10.1109/TGRS.2020.3010840

• S. Dey, A. Bhattacharya, D. Ratha, D. Mandal, H. McNairn, J.M. Lopez-Sanchez, and Y.S. Rao, 2020. "Novel clustering schemes for full and compact polarimetric SAR data: An application for rice phenology characterization". ISPRS Journal of Photogrammetry and Remote Sensing, 169, pp.135-151, https://doi.org/10.1016/j.isprsjprs.2020.09.010.