Research On Remote Sensing Image Segmentation Of Water Based On Adaptive Morphology

Surface water is an important part of water resources.Quantitative statistics of surface water is an effective means to analyze the basic situation of water resources.As a traditional water area measurement method,manual field survey not only takes time and effort,but also has large errors.However,with the successful implementation of the high-resolution project,the high-resolution satellites have been launched.The Chinese remote sensing satellites have entered the sub-meter high-resolution era.Remote sensing image of water can be analyzed by using digital image processing method,which overcomes the problem of time-consuming and large error in traditional manual water resources measurement.Because the shape,size and number of the target in the remote sensing image are different,this paper uses the adaptive morphology theory to segment the surface water area of remote sensing image.The main work of the paper is as follows:(1)The research status of surface water segmentation in remote sensing images at domestic and foreign is analyzed,and the basic theory and methods of mathematical morphology in remote sensing images are analyzed.(2)Aiming at the problem of low contrast remote sensing image,the target water and non-target ground objects have similar grayscale values,a morphological remote sensing water image logarithmic enhancement method based on adaptive elliptical structuring elements is proposed to enhance the contrast of remote sensing images while filtering the noise outside the water and maintaining the water edge.Firstly,the linear structure tensor is used to estimate the image eigenvalue and eigenvector,and the adaptive elliptical structuring elements are established to keep the edge of water better;then the adaptive morphological erosion and dilation are defined,and the opening and closing are combined;finally,the logarithmic enhancement operations is constructed by using the adaptive morphological operations,so as to filter the non target area noise and enhance the remote sensing water image.The experimental results show that compared with traditional and adaptive morphological enhancement methods.It not only filters the noise outside the target,enhances the target water,but the edge information of the water is not destroyed.The average error of image segmentation after enhancement using gray slices is less than 3.29%.(3)When using morphology to segment image of water,the fixed structuring elements will cause the change of water feature information in the image(such as generating new artificial targets,changing the edge position of larger targets,destroying the boundary of the transition zone between the original targets,and losing small targets etc.),which will affect the accuracy of subsequent water segmentation.In order to segment surface water in remote sensing image accurately,a method of water segmentation in remote sensing image usingmorphological adaptive elliptical structuring elements is proposed.The corresponding closing operation is derived to eliminate the influence of dark detail noise on the water without over stretching the water edge,so that the water edge can be maintained more accurately.On this basis,the water area can be segmented by gray slice.The experimental results show that the proposed method has higher segmentation accuracy and the average segmentation error is less than 1.43%.