基于Krisch算子和绝对关联度的图像边缘检测

Image Edge Detection Based on Krisch Operator and Absolute Correlation Degree

提出一种基于Krisch算子和绝对关联度的图像边缘检测算法。由于绝对关联度进行图像边缘检测时没有考虑到边缘方向而导致检测出的边缘不完整,因此论文首先根据绝对关联度的思想将Krisch算子中的八个方向模板作为绝对关联度中的参考序列,然后分别求出由图像像素点及其八邻域组成的比较序列和参考序列的八个关联度,计算出关联度的均值,并和设定的阈值比较大小来判定像素点是否为边缘点。实验结果表明,该算法较传统的边缘检测算法能够准确提取图像的边缘轮廓信息,设计简单、易实现。

An image edge detection algorithm based on Krisch operator and absolute correlation degree is presented in this paper.when using absolute correlation degree to detect image edge without considering direction of the image edges,which leads to detecting image edges be incomplete.So the eight direction templates of the Krisch operator are made as reference sequence of absolute correlation degree according to its thought,then respectively obtain eight correlation degrees of reference sequence composed of image pixel with eight neighborhood pixels and comparison sequence are obtained respectively,the mean of these correlation degrees is worked out,and compared with the given threshold value to distinguish whether the pixel is edge point.The experimental results show that the algorithm can accurately extract the image edge profile information,the design is simple and easy to implement,compared with traditional edge detection algorithms.