SAR image adaptive MAP filtering based on the generalized Gaussian model

Abstract

Synthetic aperture radar (SAR) images are inherently affected by a signal dependent noise known as speckle, which is due to the radar wave coherence. In this paper, we propose a novel adaptive despeckling filter and derive a maximum a posteriori (MAP) estimator for the radar cross section (RCS). We first employ a logarithmic transformation to change the multiplicative speckle noise into additive noise. We model the RCS using the recently introduced Generalized Gaussian density function [1] , Which was proved to be the best described for the SAR Amplitude. We estimate model parameters from noisy observations by means of second-kind statistics theory, which relies on the Mellin transform. Finally, we compare the proposed algorithm with several classical speckle filters applied on actual SAR images. Experimental results show that the MAP filter based on the Generalized Gaussian prior for the RCS is among the best for speckle removal. INTRODUCTIONIn the past ten years, there has been a growing interest in synthetic aperture radar imaging. SAR systems are capable of producing high-quality pictures of the earth's surface while avoiding some of the shortcoming of other form of remote imaging systems. For example, SAR imaging systems overcome the night-time limitations of optical cameras and cloudcover limitations of infrared imagers. So, it play an important role in applications such as remote sensing for mapping, search and rescue, mine detection, and target recognition.Unfortunately, SAR images are inherently affected by a signal dependent noise known as speckle, which is due to the radar wave coherence. The granular appearance of speckle noise makes it very difficult to visually and automatically interpret SAR data. Therefore, speckle reduction is a prerequisite for postprocessing SAR images.Many adaptive filters for SAR image denoising have been proposed in the past. The simplest approaches to speckle reduction are based on temporal averaging[2],median filtering, and Wiener filtering. The classical Wiener filter, which utilizes the second order statistics of the Fourier decomposition, is not adequate for removing speckle since it is designed mainly for additive noise suppression. To address the multiplicative nature of speckle noise, Jain developed a homomorophic approach, which by taking the logarithm of the image, converts the multiplicative into additive noise, and consequently applies the Wiener filter [3]. The Frost filter was designed as an adaptive wiener filter that assumed an autoregressive (AR) exponential model for the scene reflectivity [4].Kuan considered a multiplicative speckle model and designed a linear filter based on the minimum mean square error (MMSE)criterion ,optimal when both the scene and the detected intensities are Gaussian distributed [5]. The Lee MMSE filter was a particular case of the Kuan filter based on a linear approximation made for the multiplicative noise model [6]. A two dimensional Kalman filter was developed by Sadjadi and Bannnour under the modeling of the image as a Markov f...