基于Sobel算子和灰色关联分析的图像边缘检测

Image Edge Detection Based on Sobel Operator and Gray Correlation Analysis

针对灰色关联分析方法提取的边缘存在不连续、噪声敏感性等问题,本文将Sobel算子与灰色关联分析方法相结合,提出了一种基于邓氏关联度图像边缘检测的改进型算法,其融合了Sobel八方向算子技术,同时采用中值法对灰色关联度中的参考序列和比较序列实现单位一致性,并对传统的灰色邓氏关联度模型做了改进,用其求得的关联度提取图像边缘。仿真实验表明,本文算法提取的边缘连续、清晰,并在一定程度上抑制了噪声的干扰,提高了算法的效率。

Edge extracted with the gray correlation analysis method have edge discontinuities,noise sensitivity and other problems.This paper combines Sobel operator with the method of gray correlation analysis to detect edge and presents an improved edge detection algorithm based on Deng’s correlation degree that combines with Sobel eight direction operators technology and,at the same time,uses median method to achieve unit consistency for the reference sequence and comparison sequence of gray correlation degree,so as to improve the traditional gray Deng’s correlation degree,with which correlation degree can be obtained to extract image edges.The simulation results show that the edges extracted with the proposed algorithm are continuous,clear and suppressing the interference of noise to a certain extent,thus improving efficiency of the algorithm.