Canny Operator Edge Detection Of Research And Improvement

The edge detection of digital image processing is the foundation of the image analysisarea, which includes image segmentation, target area recognition regional shapeextraction and so on, and it is also a important property of the image feature extraction.Now it has become one of the most active subjects in machine vision research field.This paper firstly introduces the concept of digital image processing, and summarizesthe basic concepts and the present situation of edge detection technology. Then, thepaper analyzes the classic edge detection operator and the author makes experiments forcomparison. The paper studies Canny operator a lot which is the most optimal operator,makes a detailed description of the realization process of it, and finds some deficienciesof it. The paper proposes three improvements to the shortcoming of. Cannyoperator.One is using median filtering instead of Gaussian filtering to remove theimage noise. The other one is using OTSU proposed by OTSU to get the mostreasonable threshold of image, and then take the process of double thresholdsegmentation. In the experiment, the paper makes a comparison of different noisepollution, and the experimental result shows that the improved Canny Operatorperformances better in anti-noise process, and using the double threshold which isdefined by OTSU to take segmentation can get more accurate edge information andexpected result.