Positioning And Tracking Of Seismic Image Boundary Based On PCNN

As a modern neural network method,PCNN can group pixels with similar gray and space in image processing,which is very suitable for image edge detection and location.When PCNN is used in image edge detection,the detection effect depends on the setting of various parameters in the model,including the selection of time attenuation coefficient,link strength coefficient,threshold amplification coefficient and other parameters.At present,the parameters of PCNN model are mainly set by experience,which does not make full use of the feature information of the image itself and is not universal in image processing.In this paper,by analyzing the parameter function of PCNN model and the gray space moment relationship of image edge sub-pixel,and according to the seismic image information,the parameters of PCNN model are set adaptively to make the edge detection and location of seismic image more accurate.Due to the random convergence of the traditional PCNN model iteration,by analyzing the optimization strategy of ant colony algorithm,the image cross entropy is used as the fitness function,and combined with the image output entropy to judge the termination condition of the model iteration,so as to reduce the model iteration redundancy and improve the efficiency of image edge detection and location.In this paper,tracking the edge of the image based on the edge detection and location of the seismic image.The traditional boundary tracking method will have the problems of repeated tracking and missing tracking in the process of tracking.By analyzing the data structure of image pixels,proposing a four neighborhood image boundary tracking method based on M connectivity,and use the positive and negative sequence of the image to track the intersection complementary characteristics,in order to avoid boundary overlap and boundary loss in the process of tracking,the contour curve with continuous and closed image boundary can be obtained,the Freeman chain code of tracking sequence can be output,and the inner and outer boundary tracking of seismic image boundary can be realized according to the cross configuration of image pixel topology.