Abstract
In this letter, a texture classification-based nonlocal means polarimetric SAR (NLM PolSAR) filter is introduced and named as texture classification-based filter (TBF). In this process, a classification algorithm that identifies the data into textural variations and heterogeneity due to speckle noise is presented. Also, a similarity metric is derived to estimate the patch similarity of K-distributed covariance matrices. Hence, a filter is proposed that processes the classified data suitably using NLM patch-based filters with either K- or Wishart distribution patch similarity metrics. The filtering performance of TBF is being analyzed on full-pol single-look RADARSAT-2 and four-look Airborne Synthetic Aperture Radar (AIRSAR) data.
Accepted Version
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call