Based Multifractal Image Edge Extraction

The edge of image is the most basic features of image, the edge detection of image is one of the important foundation content for image processing , so the edge detection of image has been the hot focus in the image processing research. As one of the basic characteristics of images, The edge of image often has been used in the image high-level characteristics description, such as image recognition, image compression.Multi-fractal fractal is the further development of fractal theory, it used to describe the scaling properties of fractal geometry in the growth process of the characteristics of different levels. Every different level have different characteristic parameters that constitute a different collection of parameters or spectral function, multi-fractal used to describe Singular probability distribution which depends on fractal dimension can not be fully described, Multifractal singular spectrum function and H(o|¨)lder index can be a better description of the local characteristics of image information, the method of multi-fractal analysis is from the local departure of images to study the ultimate overall characteristics, as a emerging discipline, Multifractal has begun to apply to many areas of natural sciences, and has become one of the focus on study of non-linear science and computer intelligence.The main job of the thesis is giving a new method local of the edge detection in different measure. Simulation results show that the edge detection Based on the multi-fractal is a common local image edge detection methods, when the image has obvious fractal features, this edge detection method will produce better results, in accordance with the frequency of certain texture, given the local principal edge information ,at the same time decrease any unimportant local edge information, and reach a more satisfactory effect of local edge detection. Experiments show that the method can achieve better effect of image edge extraction, and have good validity and achievability.