Coastline Detection Of SAR Images Based On Geometric Active Contour Model

Coastline detection of synthetic aperture radar(Synthetic Aperture Radar,SAR)images plays an important role in coastline management,map automatic navigation,ship target recognition,and so on.Geometric active contour(Geometric Active,Contour,GAC)is the active contour model(Active Contour Model ACM,also ca lled Snake model)development model based on the Snake model is to extract edge areas of major breakthroughs,and research value very useful.In recent years,with the extensive research of Snake model,the idea of GAC model has received extensive attention from all over the world,and it has also become more and more widely involved.The GAC model also shows strong practic ability in extracting the boundaries of SAR images.But because of the fuzzy boundary,contrast and gray level and reduce the noise interference problem of SAR image with the method of GAC model,SAR image processing will still encounter some weak boundary problems,the number of iterations and iteration time is easily affected by the initial contour image and image preprocessing to extract SAR image coastline impact the problem.To solve this problem,the coastline detection in SAR images as the application background,the weak boundary problem,involving the initial contour image iterative coastline detection and the number of iteration time and image preprocessing affect the detection of coas tline were studied.After the study of SAR image coastline detect ion of weak boundary,puts forward the improved symbol pressure function combined with regional information for the GAC model and the boundary conditions to stop the accurate extraction of the coastline,which could make up the coast SAR image line weak boundary problem,the extracted coastline is more accurate.Based on the study of the weak boundary of the coastline of SAR images,the influence of the initial contour of SAR image and the image preprocessing on the extracted coastline is also studied in this paper.This paper will study the influence of the initial contours of different sizes of SAR images on the number of ite rations and the iterative time of coastline detection,which is based on the insensitivity of GAC model to the initial contours of images.The bigger the initial contour of the image is,the less the number of iterations of the GAC model and the fewer iteration s,the shorter the iteration time.In image processing,because of the speckle noise in SAR images is multiplicative,so the general methods of image enhancement,noise removal method has not been applied to the SAR image,this paper used gray-scale transformation method for SAR image enhancement processing,increase the contrast of the image,the filtering of the SAR image with Lee filter wave.The experimental results show that the SAR image processed by this image preprocessing method can achieve very good detection results.Test data show that this method not only can detect the SAR image coastline effectively,and compared with other related coastline detection method,which reduces the number of iterations and iteration time is shortened and the detection accuracy has been further improved,showing the validity of the method.