Speckle Noise Modeling And Reducing Technique Of Synthetic Aperture Radar Images

Automatic interpretation of the information contained in the reflected intensity of the SAR data is extremely difficult. These difficulties are due to the speckle phenomenon that can be regarded as a strong multiplicative noise affecting all coherent imaging systems. Most edge detection, target classification and area segmentation methods developed for optical images perform poorly when applied to 'SAR images. We detailed analyze the speckle reduction methods of SAR images in theory, and validate them by experiment. The main work contains these three parts:First, We study the physical mechanism of speckle and the scatter model thoroughly. Through calculation methods, We establish the method to choose different kinds of speckle model according to the difference of scale. Further more, We give estimate method to the parameters. It shows better result.Second, We study many kinds of existing speckle reducing methods: methods based on moving windows; methods based on characteristic detection; methods based on mathematical morphological operators and methods based on wavelet transform. By calculation and theory analysis. We find advantages and disadvantages of all methods mentioned above. Besides, We propose a SAR speckle reducing methods based on anisotropic diffusion. This method utilizes the property of Coefficient of Variation. And the optimal result can be obtained through iteration. This method obtains good results.At last, We study subjective criterions and impersonality criterions of speckle reducing.