Comparison of speckle filtering methods for POLSAR analysis of earthquake damaged areas

Speckle noise makes the synthetic aperture radar (SAR) image interpretation difficult and reduces the accuracy of the rapid disaster assessment. This paper introduces the influence of speckle filtering methods on the polarimetric SAR data interpretation of damages. Four types of speckle filtering methods, such as boxcar filter, refined Lee filter, Lee sigma filter, and sequential filter, are investigated to determine the best filtering options for detection of damaged areas. As a result, Lee sigma filter exhibits the best performance in the analysis of the damaged areas. The sequential filtering approach also shows a good performance, which is comparable with that of Lee sigma filter, with minimum computational cost.