图像边缘检测算法的比较与实现  

Comparison and implementation of image edge detection algorithm

作　　者：李娅娅[1] 李志洁[1] 郑海旭[1] 王存睿[1] 

机构地区：[1]大连民族学院计算机科学与工程学院,辽宁大连116600

出　　处：《计算机工程与设计》2010年第9期1971-1975,共5页Computer Engineering and Design

基　　金：大连民族学院博士启动基金项目(20086205)

摘　　要：边缘检测的目的是标识数字图像中灰度变化明显的点,而导函数正好能反映图像灰度变化的显著程度,因而许多方法利用导数来检测边缘。针对典型的基于导数的边缘检测算法Sobel、Log和Canny,在分析其算法思想和流程的基础上,利用MATLAB对这3种算法进行了仿真实验,分析了各自的性能和算法特点,比较边缘检测效果并给出了各自的适用范围。The purpose of edge detection is to mark the image points that change significantly in intensity. Since derivative could reflect the significant degree of image intensity changes, derivative is used to solve the problem of edge detection in many researched. Three typical edge detection algorithms based on derivatives （Sobel, Log, Canny） are analyzed thoroughly in aspects of idea and flow. Experiments are performed using MATLAB to evaluate these edge detection algorithms. The performance results are analyzed and compared with the scope of application respectively.

关 键 词：图像处理 边缘检测 模板 卷积 非极大抑制 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]