ABSTRACT:
In this project, we propose another test measurement for unsupervised change discovery in polarimetric radar pictures. We work with multilook complex covariance network information, whose fundamental model is thought to be the scaled complex Wishart appropriation. We utilize the perplexing kind Hotelling-Lawley follow (HLT) measurement for estimating the likeness of two covariance networks. The conveyance of the HLT measurement is approximated by a Fisher-Snedecor appropriation, or, in other words characterize the noteworthiness level of a false alert rate managed change indicator. Analyses on mimicked and genuine PolSAR informational collections exhibit that the proposed change location technique gives discovery rates and mistake rates that are tantamount with the summed up probability proportion test.
BASE PAPER: Detecting Changes in Fully Polarimetric SAR
