Despeckling Algorithm Of SAR Images Based On Structure Information Detection And Fisher Model

Synthetic Aperture Radar(SAR) is a kind of active microwave remote sensor, which could use pulse compression technique to promote range resolution as well as using the principle of Synthetic Aperture to enhance the azimuth resolution, so that it can obtain high resolution SAR images. SAR has so many characteristics, such as all-time, all-weather, high resolution and the penetration capabilities. As to the above characteristics, the SAR has many important applications value in the military, agriculture, marine, ecology and geography. SAR images are obtained by coherent wave radiation. Owing to the coherence of radar waves, it is inevitable to produce some speckle while imaging processing. The speckle play a main role in affecting SAR image quality, so it can degrade the quality of SAR image, and affect the subsequent edge detection, image segmentation, object recognition and image processing. Therefore, it is an indispensable issue to be solved that how to reduce speckle effectively in SAR image processing. In this paper, further research had been discussed about how to suppress speckle effectively in SAR image. The accomplished work as following.(1) A modified Frost filtering algorithm with adaptive windowing and adaptive tuning factor had been proposed. In traditional Frost filtering algorithm, adjusting factor and the windowing size would influence the performance of Frost filtering significantly. When adjusting factor and windowing are too big or too small, they could not get an effective balance between speckle suppression and edge preservation. Nevertheless, in this paper, through the adaptive changing of the windowing and adaptive factors, the proposed algorithm can not only suppress speckle effectively, but also keep the marginal information of image better.(2) The Fisher distribution model had been introduced, and several kinds of parameters estimation methods about the Fisher distribution had been given at the same time. The unknown parameters in the Fisher distribution could be estimated through the appropriate parameter estimation method. Then the high resolution SAR image modeling can be built based on the Fisher distribution model. Through modeling experiments, we can know that the statistic characteristics of the high resolution SAR image peak and heavy-tailed could be described better through the Fisher distribution model.(3) Based on the second-kind statistics and Mellin transform, MAP filtering equation based on the Fisher distribution had been derived with the help of the maximum a posteriori probability(MAP) criterion. Owing to the different areas features in SAR image had not been taken into account, it cannot suppress speckle in SAR image effectively. Therefore, Fisher MAP despeckling algorithm of SAR images based on the local structure information detection had been put forward in this paper. It can distinguish the point target, edge areas and uniform areas in the image effectively, and could adopt different despeckling algorithm methods according to different areas. Through the despeckling experiments, it could validate that the proposed algorithms can suppress speckle in homogeneous area fully, while it can preserve the edge information and point targets effectively. Therefore, the proposed algorithm can get a better balance between speckle suppression and edge preservation.