Study On Method Of Remote Sensing And Medical Image Segmentation Based On Active Contour Model

Image segmentation is one of the most important tasks in image processing.In recent years,the image segmentation method of partial differential equations has been developed rapidly because it is based on numerical calculation in solving and other advantages,but there is no unified PDE segmentation method of image segmentation based on all types of.This paper mainly focuses on the remote sensing image and medical image segmentation based on active contour model.The remote sensing image segmentation is the main purpose of extraction in remote sensing image "interest" of useful information,such as weather,land,sea and other key information to enhance mutual understanding,and according to the information of some forecasts of the future;and the main purpose of medical image segmentation is to help the doctor for the patient's lesions to do may the accurate segmentation,and the patient's condition to make accurate judgments,which not only reduces the workload of doctors and patients for follow-up treatment provide precious time.This paper first introduces the research status of image segmentation,and mainly introduces the related mathematical theory of image segmentation based on partial differential equations;in the traditional model of the problem such as high computational complexity,and sensitive to the initial contour segmentation efficiency is studied,put forward corresponding solutions.The innovative work of this paper lies in the following aspects:1.A remote sensing image segmentation algorithm based on the active contour model based on the non sampling contourlet transform(Nonsampling Contourlet Transformation,NSCT)is proposed.Research on statistical characteristics of NSCT at first directional subband coefficients,analyzed the relationship between NSCT coefficients and marginal statistics subbands and joint statistics characteristics,obtained the corresponding statistics;on this basis,establish a segmentation model of vector CV image based on NSCT domain.The NSCT can take into account the detail information of the image,and image region global information CV model as the model's energy function,can get the overall information of the image,in the image region of interest or target segmentation.2.A combination of GAC model and SBGFRLS(Selective and Filtering Regularized Level Set Gaussian,SBGFRLS)Binary model for active contour segmentation of medical images.First of all,by weight function based on the model,the possibility of weight function is designed to improve the zero level set inspection of medical image contours,to further simplify the model level set algorithm,introduced Gauss filtering.At the same time,in order to improve the convergence speed of the proposed segmentation model,this paper introduces the Hausdorff distance measuring distance of target image,to solve the traditional model of medical image segmentation in convex problems.This model improves the ability of capturing the edge of multi object in medical images,and has a high accuracy.Image segmentation is an important part of image processing,which has a great effect on the subsequent processing.In this paper,using the active contour model segmentation,combined with the characteristics of remote sensing images and medical images,remote sensing image segmentation algorithm were designed NSCT coefficients and Hausdorff distance and medical image segmentation based on the extracted target can be used for other applications such as object recognition,remote sensing image medical image positioning the tumor region.