基于Matlab图像边缘检测算法效果对比  

Study on Edge Detection Algorithm in Image Processing

作　　者：杨先花[1] 黎粤华[1] 

机构地区：[1]东北林业大学,黑龙江哈尔滨150040

出　　处：《机电产品开发与创新》2010年第2期129-131,共3页Development & Innovation of Machinery & Electrical Products

摘　　要：边缘是图像最基本的特征,包含图像中用于识别的有用信息,边缘检测是数字图像处理中基础而又重要的课题。文章具体考察了5种经典常用的边缘检测算子以及最小二乘支持向量机提取边缘检测算子,并运用Matlab进行图像处理结果比较。梯度算子简单有效,LOG算法和Canny边缘检测器能产生较细的边缘。最小二乘支持向量机结合图像的梯度和零交叉信息,选取一定的参数条件,可以获得比Canny方法更好的性能。Edge is the most basic features of the image, including images which are used to identify useful information, edge detection is the basical and important issue of digital image processing. The article specifically examines five kinds of classical edge detection operator which are commonly used and least squares support vector machines which extract edge detection operator, and the image processing results are compared by Matlab. Gradient operator is simple and effective, LOG algorithm and Canny edge detector can produce smaller edge. Least squares support vector machine which combine image of the gradient and zero-crossing information can get better performance than the Canny method by selecting a certain parameter conditions.

关 键 词：PID神经网络 BP算法 发动机油门控制 

分 类 号：TP23[自动化与计算机技术—检测技术与自动化装置]