An Adaptive Clustering MAP Algorithm to Filter Speckle in Multilook SAR Images
F.N.S. Medeiros, Nelson Delfino Davila Mascarenhas, Luciano da Fontoura Costa · Biblioteca Digital da Memória Científica do INPE (National Institute for Space Research) · 1998
Abstract. SAR images are generally affected by a multiplicative noise, called speckle, which degrades the quality of these images. Using this model we present an algorithm based on the Maximum a Posteriori (MAP) criterion to reduce speckle noise of multilook amplitude data. The speckle in these images is approximately described by the Square Root of Gamma distribution, which is used to develop MAP filters using different a priori distributions. We also suggest the combined use of MAP and the kmeans clustering algorithm as a formal way to choose the best window size to update noisy pixels. We conclude this work by calculating the coefficient of variation, defined as the ratio of the standard deviation to mean, of MAP filtered images and of the original image to measure the reduction of the speckle in homogeneous areas.