Synthetic Aperture Radar Image Speckle Suppression And Edge Detection Algorithm

Synthetic aperture radar(SAR)is a high-resolution microwave remote sensing imaging radar with all-weather and all-day advantages.SAR image is the main product of Synthetic Aperture Radar(SAR),which contains a large number of speckle noises with multiplicative features.The existence of speckles has caused the SAR image's automatic interpretation processing to be greatly affected.Therefore,the research on speckle suppression in SAR images is of great significance.Edge detection is another important link in SAR image interpretation processing.Its research also has important significance.This paper focuses on the study of SAR image speckle suppression.A non-local average SAR image speckle suppression algorithm is proposed and its improvement is discussed.At the same time,this paper also preliminarily studies the SAR image edge detection algorithm,and proposes an SAR image edge detection algorithm.In this paper,a non-local mean speckle suppression algorithm based on the mean ratio and the coefficient of variation is proposed.The algorithm first calculates the mean ratio of the center pixel and other pixels in the search window using the similarity window as a measure of similarity,thereby replacing the Gaussian Weighting Euclidean distance used to measure the similarity in the traditional non-local averaging algorithm.;then calculating the coefficient variation of the entire image to construct an adaptive attenuation factor,thereby improving the constant attenuation factor of the conventional non-local averaging algorithm,making it to the SAR image different regions are adaptively filtered;finally,the new similarity metrics and attenuation factors are combined with a negative exponential weighted model to construct a filter function,and the image is non-locally average filtered.In addition,this paper continues to improve the proposed algorithm for suppressing plaque from three directions and proposes three improvements.The simulation experiment was carried out on the new patching algorithm and its improved algorithm.The performance of the algorithm was greatly improved in both visual effects and parameter indexes.Classical differential edge detection operator has been proved to be not directly applicable to SAR images interfered by multiplicative speckle noise.For this reason,a SAR image edge detection algorithm based on ratio operation and anisotropic Gaussian kernel is proposed.The algorithm generates a new edge intensity map by multiplying the ratio calculation and the edge intensity map calculated by the anisotropic Gaussian kernel.Then the edge detection image of the SAR image is obtained by comparing the equalization,non-maximum suppression,and double threshold decision.Simulation experiments are performed on the proposed new edge detection algorithm.The resulting SAR image edge image is more visually improved than the traditional ratio calculation algorithm.This paper first introduces the background significance and imaging principle of SAR image research,classic image speckle suppression algorithm for SAR images,and classic edge detection operator.Then a SAR image speckle suppression algorithm and its series of improved algorithms are proposed.Finally,an edge detection algorithm for SAR image is proposed and summarizes and looks forward to the full text.