基于尺度空间边缘强度累计的边缘检测  

Edge Detection Based on Edge Intensity Accumulation in Scale-Space

作　　者：胡跃明[1] 陈泽武[1] 陈安[1] 高红霞[1] 

机构地区：[1]华南理工大学精密电子制造装备教育部工程研究中心∥自动化科学与工程学院,广东广州510640

出　　处：《华南理工大学学报（自然科学版）》2014年第9期59-64,共6页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家"863"计划项目(2012AA041312);国家科技重大专项(2014ZX02503);国家自然科学基金资助项目(61105081);广东省新兴战略性产业重大专项(2010A081002007)

摘　　要：为抑制图像中的细微结构、噪声、弱边缘等干扰对边缘检测的影响,同时提高边缘检测的位置精度,提出了一种基于尺度空间中边缘强度累计特性的边缘检测方法.该方法首先通过比较目标对象在多个尺度下的平均边缘强度来选择检测目标边缘的最佳尺度,然后依据边缘点在尺度空间上的强度累计值的差异滤除该尺度下的干扰边缘.实验结果表明,文中方法能够在少量的平滑预处理条件下有效地抑制上述各种干扰,提高图像中目标边缘的稳定性与位置精度.In order to reduce such effects of interference as microscopic structure,noise and weak edge on the image edge detection and to improve the positional accuracy of the detection,an edge detection method is proposed on the basis of edge intensity accumulation characteristic in scale-space.This method selects the optimal scale of target edges by comparing the average edge intensity of target object at different scales,and then filters out all the jamming edges at the optimal scale according to the difference of edge intensity accumulation values of edge points in scale-space.Experimental results show that the proposed method can effectively restrain all the above interferences with slight smoothing,and thus the stability and positional accuracy of target edges is improved.

关 键 词：边缘 边缘强度 尺度 累计 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]