Image Edge Detection Method Based On Mathematical Morphology

Image edge is one of the most basic characteristics of the image, and edgedetection is an important link of image pre-processing and analysis, with a widerange of theoretical and practical significance. Compared with traditional edgedetection methods, image edge detection based on mathematical morphology isbetter, since it can change the scale of the morphological structuring element toovercome the effect of noise, and change the structure and orientation of structuringelement to detect richer edges. This method can not only meet the real-timerequirements, but also be easily to implemented in hardware.This paper first introduces the basic concepts and development status ofmathematical morphology and edge detection, elaborates traditional edge detectionmethods, performs the experimental analysis of the image with and without noise,simultaneously, and then introduces some new emerging edge detection methodsbriefly. Since the method of edge detection in this paper is that based onmathematical morphology, we then introduce the basic theory of mathematicalmorphology, present some basic morphological operator formulas, and simulate inthe case with and without noise and carry out a detailed analysis and summary.According to the characteristics of morphological operators and the disadvantagesof selecting single structuring elements, this paper gives two improvedmorphological edge detection operators.The first improved operator is the formula constructed based on thecharacteristics of erosion, dilation, opening and closing operations to. In the imageswith noise, by using the open-close operation of this formula repeatedly to filter,we can effectively suppress noise. At the same time, since the selected structuringelements have the characteristic of multi-directional, this method can detect theedge information in different directions to ensure the integrity of the edgeinformation.The second improved operator uses different structuring elements on the basisof the first improved operator, and has the characteristics of multi-structure,multi-scale and multi-direction. The multi-structuring element can detect varioustypes of image edges, while rational combination of the multi-scale structuringelement can suppress noise effectively, and simultaneously detect better edgedetails.Finally, this paper experiments with the two improved operators above, andcompares with the previous morphological operators and traditional operators. It can be seen from the results that the improved operators are better on the effect ofedge detection and denoising performance, and can be used widely in the laterimage processing.