Research On Application Of Multiscale Geometric Analysis Technology For SAR Image Processing

Synthetic Aperture Radar(SAR) is a kind of high resolution imaging radar, which has been developed since World WarⅡ. High-resolution radar image can be obtained under all weather, day-and-night and long distance conditions, so it is applicated abroad both in military and civilian fileds. The Research on algorithms of SAR image processing including speckle suppression and target detection has important theory significance and application value. Curvelet transform and Contourlet transform are two new Multiresolution Geometry Analysis(MGA) methods which can represent images most effectively. Combined with the the Natural Science Fund of Jiangsu Province(N0.BK2001047) which name is"Research On SAR Image Evaluation", this paper uses above-mentioned methods to SAR image processing, proposes new algorithms of speckle suppression and target detection for SAR image and researches on abilities of these algorithms.This paper introduces main fields in SAR image processing and the introduction emphasizes on speckle suppression which is the main field our task researches on; This paper also introduces MGA and the introduction emphasizes on Curvelet transform and Contourlet transform.Combined with characters in SAR image, this paper proposes speckle suppression algorithms for SAR image based on Curvelet transform, Contourlet transform and SOT structure in Contourlet domain separately. The denoised images obtained by these algorithms gain pretty good visual effect and all indices are enhanced effectively comparing to the other denoised images. These denoising algorithms protect target's texture and edge characters efficiently while suppressing speckles successfully.This paper utilizes the best representation ability for 2-dimentional singular signal of Contourlet transform, proposes a detection algorithm for airfield target of SAR image based on Contourlet transform. The detection result shows higher capability in shape and edge of airplane protection than the result which is gained using wavelet transform.The tag of this paper makes a summary of research and gives expectation on research future about applications of MGA in SAR image processing.