Research On Image Edge Detection Based On Gradient Operators

The image edge detection is one of the most important technologies in the digital image precessing and computer vision, playing an important role in image processing system. It is a critical factor affecting system’s performance. The performance of computer vision system is influenced directly by the accuracy of the edge detection algorithm. Most classic edge detection algorithm, such as gradient operator, Laplace operator, LOG operator, only detect edge of restricted types and they are sensitive to noise, too. It is a difficult topic in image processing research field to detect the edge correspond to the real boundary of target in image.This paper researches the edge detection algorithm based on gradient operator and its problems, such as lower detection accuracy and noise resistance ability. An improved Sobel image edge detection algorithm and an image edge detection algorithm based Simulated Annealing and gradient operator are introduced. Sobel edge detection algorithm is influenced seriously by noise and can only detect edge of restricted types. According to these problems, the improved Sobel image edge detection algorithm add noise detection and edge thin algorithm. The image edge detection algorithm based Simulated Annealing and gradient operator use Simulated annealing to get the best operator,which is used to detect edge. This solved the problem of using the settled operator.The experiments proved that the improved Sobel edge detection algorithm can detect edge of different types. It have a high accuracy and good noise resistance ability. The edge detection algorithm based on Simulated Annealing and gradient operatorhas a good detection result, high statility and good noise resistance ability.In the further research work, the efficiency of the two algorithm and the edge type they adapted to will be researched..