In order to detect the image edge effectively, a new algorithm combining the general grey correlation with the LoG operator is proposed in this paper, which determines whether the discrete sequences have a close relationship or not in terms of the degree of similarity in trends of their curves. Firstly, a 3×3 pixel template is created in the grey image being detected. The center pixel and other eight pixels around it are expanded to make up the behavioral sequence, and four LoG operators are chosen and expanded in the same way to be the reference sequence. Next, the general grey correlation degree between the behavioral sequence and the reference sequence is calculated. After an appropriate threshold is set, the image edge can be detected out by comparing the grey correlation degree and the threshold. Experimental results show that this method can detect the edge effectively with higher accuracy.
[1]
Dai Wenzhan,et al.
On Image Edge Detection Based on Grey Absolute Degree of Incidence and LOG Operator
,
2006,
2007 Chinese Control Conference.
[2]
Yan Jie.
An Image Edge Detection Algorithm Based on Grey Relational Analysis and Canny Operator
,
2006
.
[3]
Zheng Zihua,et al.
An Algorithm of Image Edge Detection Based on Gray Weighted Absolute Correlation Analysis
,
2006
.
[4]
Dai Wen-zhan.
Approach of image edge detection based on Sobel operators and grey relation
,
2006
.
[5]
Yi-xin Yin,et al.
A CNN-based edge detection algorithm for remote sensing image
,
2008,
2008 Chinese Control and Decision Conference.
[6]
LI Xin-wu,et al.
A Novel Algorithm of Image Edge Detection Based on Gray System Theory
,
2003
.