改进的基于Canny算子的图像边缘提取方法精要.docVIP

摘 要 边缘检测主要是图像的灰度变化的度量、检测和定位，其实质就是提取图像中不连续的特征，因此边缘检测是图像分割领域的一部分。边缘检测是数字图像分析的前提，检测结果的优劣影响着下一步图像压缩、计算机视觉、模式识别等的应用。因此对它的研究具有现实意义和理论意义，在工程应用占有十分重要的地位。 本文对边缘检测理论和算法作了深入的研究，在具体分析各类传统的边缘检测算法的基础上，这里重点研究Canny算子边缘检测算法，并结合传统的Canny算法的双门限选取方法进行改进，达到自适应阈值的效果。最后，利用MATLAB软件工具实现该算法。实验结果表明，改进后的算法较传统的Canny算法取得更好的边缘检测效果。 关键词：图像处理；边缘检测；Canny算子；自适应阈值 Abstract Edge detection is mainly aimed at measurement, detection and location of the change of image intensity,its essence is to extract the characteristic of the discontinuous parts of images. So the edge detection is part of the image segmentation. Edge detection is the premise of digital image procession, its detection result plays an important impression on the next processing, such as image compression, computer vision, pattern recognition, and so on. Hence, the study of it has practical significance and theoretical significance, and it occupies a very important role in engineering applications. This paper makes an in-depth study on the theory and algorithms of edge detection. Based on specific analysis of various types of traditional edge detection algorithm, this paper focuses on the Canny algorithm.. In this paper, Combined with the traditional double threshold selection method to improve the Canny algorithm to make the thresh adaptive.the improved to refine the dual-threshold in the Canny algorithm to make the thresh adaptive. Finally, using MATLAB software tools to implement the algorithm. The experimental results show that the improved algorithm gains better edge detection results than the traditional Canny algorithm.. Key words：mage Segmentation; Edge Detection; Canny;Adaptive Threshold  目 录 第1章 概述 1 1.1 课题的研究背景及意义 1 1.2 课题的研究现状 3 1.3 本文主要研究内容 4 1.4 本章小结 5 第2章 图像的边缘检测与提取 6 2.1 边缘检测过程 6 2.2 边缘的类型 7 2.3 边缘的判定 8 2.4 本章小结 9 第3章 常见边缘检测算法 10 3.1 Roberts（罗伯特）边缘算子 10 3.2 Prewitt边缘算子 11 3.3 Sobel边缘算子 12 3.4 Laplacian边缘算子 12 3.5 LOG（高斯-拉普拉斯）边缘算子 13 3.6 本章小结 15 第4章 传统的Canny算子边缘检测 16 4.1 Canny算子的基本原理 16 4.2 Canny边缘检测算法步骤 17 4.3 Canny算子与其他算