Reveal Difference in Synthetic Aperture Radar Images Implementing Fuzzy Clustering Along With Improved MRF Energy Function and Wavelet Denoising Technique

Abstract

In this paper we put forward a novel method to reveal difference in synthetic aperture radar (SAR) images. In this approach we classify the changed and unchanged region by the help of the fuzzy c-means (FCM) clustering along with the use of an Improved Markov random field (MRF) as energy function. It is important to deal with speckle noise that is found in SAR images. So we use Improved MRF energy function along with FCM and a Wavelet denoising technique to reduce the speckle noise found in the SAR images. In this we use two methodologies for the detection of change in synthetic aperture radar (SAR) images. First, we apply the Wavelet Bayesian denoising technique to reduce the speckle noise. Then the image fusion technique is introduced to generate a difference image by using complementary information from a mean-ratio image and a log-ratio image. Later with the help of FCM and the improved MRF we detect the variation in the SAR images. The main advantage of the proposed method is its dominance in reducing speckle noises and its computational simplicity.