Adaptive Mathematical Morphology Based On The Weight Of Similarity

Mathematical morphology is widely used in various fields of image processing because of its solid theoretical foundation and simple algorithm. In the traditional mathematical morphology, the fixed structure elements generally cannot adapt to all regions of the images, therefore, selecting the suitable structural elements is vital important.This paper analyzes the adaptive gray mathematical morphology of the existing methods which selects the structure elements according to the local or global features of image. The methods cannot be self-adaptive and the thresholds are given by manpower, impacting the results of image processing seriously. This paper realizes a method of selecting structure element in adaptive gray morphology based on the concept of the weight of center pixel similarity. The algorithm has the advantages of setting up the automatic threshold and considers the local feature of image pixels and the global features of continuous pixels compared with the existing methods. The experimental results show that the new method has a good application in the gray image de-noising processing and image edge detection.This paper puts forward the concept of the weight of center color pixel similarity based on the rule of selecting structure element in gray image and the features of color image, setting up the criterion of automatic threshold selection. Moreover, the paper designs and implements a new algorithm of selecting structure element adaptively based on the weight of center color pixel similarity, and proposes a new adaptive morphology operator in color images. The new operator considers not only the local feature of image pixels but also the global features of continuous pixels, keeping the correlation of color image information and avoiding vector sort of color image processing, and has the advantages of selecting automatically threshold and structure element. The experimental results show that the new algorithm has a good application in color defects image edge detection.