Urban Area Mapping from Polarimetric SAR data using Fuzzy Inference System

摘要

In this work, we present urban area mapping from full-polarimetric synthetic aperture radar (SAR) data using fuzzy inference system (FIS). In particular, our aim is to utilize the profound knowledge available about scattering mechanism from urban targets to delineate urban environment. In this approach, we have utilized the recently developed polarimetric SAR scattering power decomposition technique (SD-Y4O) given in Bhattacharya et. al. The improved powers along with some other polarimetric parameters were used in this study. A suitable normalization procedure was adapted to handle the skewness in the estimated parameters. The fuzzy if-then rules were constructed from the in-depth knowledge of scattering mechanisms from an urban environment. Suitable methods were introduced to define the fuzzy inference system. The defuzzified membership values were thresholded using an unsupervised clustering method (k-means). The pixels lying in the range [μ_(max) - σ, μ_(max) + σ] corresponds to urban areas where μ_(max) is the largest cluster center and σ is the standard deviation of the cluster corresponding to μ_(max). The extracted urban area is in visually good agreement with the high resolution optical image. ALOS PALSAR full-polarimetric L-band SAR data has been used in this study.