图像边缘检测中的微分算子法及其比较

摘要

边缘是图像最基本的特征,边缘检测是图像分析与识别的重要环节.基于微分算子的边缘检测是目前较为常用的边缘检测方法.通过对Roberts,Sobel,Prewitttt,Canny和Log等几个微分算子的算法分析以及MATLAB仿真实验对比,结果表明,Roberts,Sobel和Prewitt算子的算法简单,但检测精度不高,Canny和Log算子的算法复杂,但检测精度较高.在应用中应根据实际情况选择不同的算子.%The edge is the most basic of characteristic of image. The edge detection is an important step of image analysis and identification. So far, the edge detection method based on differential operators has been one of the most common ways in the edge detection of image. By the algorithm analysis of several differential operators such as Roberts, Sobel.Prewitt, Canny, Log, etc. And the MATLAB experience comparison, can be resulted that Canny and Log are more complicated than Roberts, Sobel and Prewitt in algorithm, but more exact in detection. The appropriate operator is chosen in accordance with different situation in use.