Variational Level Set Model Of Implicit Surface Image Segmentation And Its Solving Algorithm

As an important part of digital image processing technology,image segmentation is not only a necessary link from the analysis and resolution of image processing problems,but also provides essential preparations for image information recognition and machine vision.It is used in safety information recognition and traffic control systems.,Satellite image positioning,medical imaging and other fields have a wide range of applications.Variational level set methods based on partial differential equations are becoming more mature and perfect.The high efficiency of this method has gradually become one of the mainstream methods of image segmentation,and its expansion and implementation in implicit surface image segmentation has also become a research difficulty and hot spots.The image segmentation variational level set Chung-Vese model uses a level set function to segment multiple regions.Compared with the traditional image segmentation method that uses several feature identification functions to mark different regions,it only needs to take the extreme value of a function,which improves Efficiency reduces the amount of calculation.By introducing intrinsic gradient and intrinsic dispersion,this paper designed a Chung-Vese model for three-dimensional surface image segmentation,combined with auxiliary variables to construct the corresponding alternating direction multiplier method(ADMM: Alternating Direction Method of Multipliers).To solve the original problem,it has good efficiency.The image processing on the implicit surface is closely related to the shape and characteristics of the surface.Traditional surface construction methods combined with level set functions and complex implicit equations make it difficult to construct equal-scale models of real mountain bodies,and the vegetation on the mountain surface will form different levels of distribution due to changes in altitude.In this paper,by using the elevation data in the remote sensing image,by setting the center point in the threedimensional matrix,selecting the average value,adjusting the scale and other operations,the mountain surface model is constructed and segmented.Not only can the real mountain area be obtained,but also different levels The area of vegetation.At the end of the paper,through numerical experiments,the efficiency and robustness of the model are verified.