1、摘 要 边缘是图像最基本的特征之一， 故图像的边缘检测是图像处理的主要内容之 一，也一直是图像测量技术研究中的热点和焦点。本文从边缘检测的“两难”问 题出发，对实际图像中可能出现的边缘类型进行了数学模型描述，并研究分析了 传统边缘检测算法的特点。介绍了各种算子边缘检测的基本原理，在此基础上， 采用传统算法对加入高斯白噪声以后的图像进行了边缘检测分析。 最后针对传统 Canny 算子在滤波过程中存在的缺陷，给出一种基于自适应平滑滤波的改进 Canny 边缘检测算子。通过对实验图像的分析表明，改进的检测算法对图像边缘 提取具有较好的检测精度和准确性，抗噪性能良好。 关键词：图像处理，边缘检测，Ca
2、nny算子，检测性能 ABSTRACT -II- ABSTRACT Edge is the most basic feature of image, therefore, the image edge detection is one of the main content for image processing, it also has been the hot issues of image measurement technology. In this paper, the “dilemma“ problem of edge detection is introuduced, and
3、the possible mathematical models of actual image edges are described, and the traditional characteristics of the edge detection algorithm are analyzed. A variety of the basic principles of edge detection operators are introduced. On this basic, using the traditional method to detect the edge of the
4、image which is added Gaussian white noise. Finally, an adaptive filter based Canny edge detector is given in order to eliminate the defects of the traditional Canny operator. Though the analysis of experimental images, improved detection of image edge detection algorithm has good precision and accuracy of detection, anti-noise performance. Key words: Image Processing,Edge Detection, Canny Operator, Detection Performance -III- 目 录 第一章 绪论 1 1.1 图像边缘检测方法的研究现状. 1 1.2 图像边缘检测方法. 2 1.3 本文研究的主要内容及安排. 3
