Research On SAR Speckle Reduction Considering Edge Preserving

Synthetic Aperture Radar(SAR)is an active microwave remote sensing system.It has become an important technical means of earth observation in the international frontier field by virtue of its incomparable all-day and all-weather earth observation ability and strong penetrating performance to some ground objects.With the rapid development of SAR system in recent decades,the application based on SAR image also has a broad prospect.At present,there are many applications for SAR imaging,which have penetrated into many fields such as environmental monitoring,global change,urban planning,resource exploration,disaster detection and assessment,and planetary exploration.However,as with other coherence imaging systems,speckles inevitably occur in SAR images.The speckle pattern is ubiquitous in SAR image,which seriously reduces the visual effect of SAR image,increases the difficulty of SAR image interpretation and processing,and greatly restricts the reliability and effectiveness of SAR image feature extraction,target tracking and other interpretation processing technologies.Therefore,the study of speckle suppression in SAR images has important theoretical and practical significance.The better speckle suppression processing usually consists of two points: firstly,the smooth performance of speckle noise;secondly,the retention ability of edge details.This paper aims to study the maintenance and optimization of detail retention performance on the basis of considering the better speckle suppression,and carries out the following related studies on this topic:(1)Combing and analysis on speckle formation mechanism and statistical model.For the purpose of exploring the essence of speckle formation in SAR image,this paper introduces the SAR coherent imaging system and imaging processing,systematically combs and describes the principle and process of speckle formation,analyzes the specific forms of speckle in real SAR image,introduces the speckle noise model,and derives the mathematical statistical model of SAR intensity and amplitude image,which provides the necessary basic theoretical research on speckle inhibition.(2)Study on reducing of speckle by anisotropic diffusion technology.The processing of anisotropic diffusion(AD)speckle reduction method is introduced systematically,and the advantages and disadvantages of traditional AD filter are analyzed in detail.Based on the limitations of the traditional ad filtering algorithm,two improved anisotropic diffusion algorithms are proposed,which are directional anisotropic diffusion speckle reduction(DCAD)and surface variational anisotropic diffusion speckle reduction(SVAD).DCAD method and SVAD method use the direction coherence and surface variation function to construct the edge detection operator in AD filter,optimize the anisotropic diffusion function,and give better consideration to edge preservation and speckle suppression.Through the experiments of simulation data and real SAR image,the filtering effect between DCAD method and SVAD method is compared and analyzed,and the effectiveness of the algorithms proposed in this paper is verified.(3)Research on edge information optimization compensation after filtering.Based on the ratio image,this paper analyzes the features of filtered image and the existing problems of filtered image,deduces the edge loss coefficient and effectively analyzes its influence on SAR filtered image,and proposes a method of edge information compensation(RCom)based on ratio image.This method obtains the edge loss coefficient by filtering the ratio image generated by the filtered image and the original image,and constructs the mathematical function model based on the edge loss coefficient and the initial filtering result,so as to obtain the final filtering result.The validity of RCom algorithm is verified by the experiments of simulation data and real SAR image.(4)Experimental verification of edge extraction by filtering algorithm.It is an important research direction to consider both speckle smoothing and edge information preserving.To some extent,the effect of SAR image interpretation has been further improved because of the effective speckle suppression processing which takes edge preserving into account.Taking edge detection as a specific application direction of SAR image interpretation,this paper discusses and verifies the effect of the speckle reduction processing method(DCAD method,SVAD method,RCom method)proposed in this paper on image edge information extraction through specific edge detection experiments.In addition,the effects of different filtering methods on SAR edge information extraction were analyzed and evaluated by comparison.