The Research Of Image Edge Detection Operator Based On Vision Theory

Edge is primary characteristic of image, and it contains most information of image. Edge detection technology is the base of the image measuring, image segmentation, image compressing and pattern recognition and so on. Edge detection is one of the important subjects in the field of computer vision and image measuring. How to detection the edge of object exactly and quickly is a resourceful hotspot of image processing.The subject of this paper is according to an item named'Research and simulation on automatic target recognition algorithm based on biological vision model',which is supported by fund. Some classic methods and efficient edge detection methods are introduced in this paper. The basic concepts of these techniques are summarized. The theory and deducing process of the method are analyzed, further more, each was tested to be compared the effect of edge detection. This article focuses on research of the edge detection operator of Laplacian of Gaussian (LOG). By mathematical deducing and importing vision parameters, namely improving by combination with biological vision anisotropic receptive field model, Gabor rotary factors and flexible receptive field, it deduces a mathematical model of edge detection algorithm, which is more accordant with human vision mentality. Furthermore, it inducts angle information as well as measurement information to the isotope LOG operator, which enable effective detection of edges on different directional. The extended LOG operator not only enhance the function of edge detection, but also preserve all the advantages of the classic LOG operator and with better image edge sampling capability. It was proved by experiment that the improved receptive field model is more accordant with human vision mentality than conventional isotope receptive field model; moreover, its edge detection effect is obvious advantage of conventional isotope anisotropic receptive field operator.