左中右差分绝对值最大法的边缘检测方法研究  

Study of Edge Detection Based on Left-Middle-Right Difference Most Absolute Value

作　　者：周秀云[1] 何晓渝[1] 冯中正[1] 

机构地区：[1]电子科技大学自动化学院,成都610054

出　　处：《电子科技大学学报》2010年第3期388-391,共4页Journal of University of Electronic Science and Technology of China

基　　金：四川省应用基础研究项目(2008JY0024-1)

摘　　要：为检测高温环境中微弱信号边缘,提出一种基于区域的左中右差分绝对值最大的边缘检测算法。该算法是图像灰度的统计信息的差分,在一个指定的包含有边缘信息的区域内,综合考虑单个的像素或其周围的像素点灰度值在一个方向上的累加,然后对左、中、右相邻3个像素或累加值两两差分,取3个差分值中绝对值最大的为"左中右差分绝对值最大法"结果。实验表明该方法是可行的,能检测常见经典边缘检测算法不能检测到的微弱信号,对边缘的检测能定位到像素级。An algorithm of left-middle-right difference based on a small area tracking arithmetic is proposed for detecting feeble edge signal in high temperature condition.In a appointed small area,CCD pixel’s gray value in vertical orientation is accumulated.The difference between any two of three sequential pixel accumulative total is obtained and the most difference absolute value is regarded as the method’s result.The experimental result shows that the mothed is feasible and can detect feeble edge which classical edge mothed can not recognise,and its precision can reach pels level.

关 键 词：边缘检测 微弱信号 图像处理 左中右差分 检测算子 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]