本科毕业设计(论文)
检测算法研究
学 院：信息工程学院 专 业：自动化 学 号： 学生姓名： 指导教师：
二○一 年 五月 二十三日
题 目：基于matlab的图像边缘
2013届毕业设计（论文）
基于matlab的图像边缘检测算法研究
摘要
图像的边缘检测技术是数字图像处理技术的基础研究内容，是物体识别的重要基础。现有边缘检测技术在抑制噪声方面有一定的局限性，在阈值参数选取方面自适应能力很差，有待进一步改进和提高。
本论文首先介绍了图像边缘检测这个课题的意义和背景；作为理论基础，在第二章简单的介绍了传统的图像边缘检测算法，如Roberts算子、Sobel算子、Prewitt算子、Laplacian算子、LOG算子，回顾了经典的边缘检测算法，为后面介绍Canny算法作为铺垫。在第三章，结合Canny算法的基本原理、算法的三个标准、算法的思路及检测步骤提出了对Canny算子中的图像滤波平滑处理及取阈值的算法进行改进的方法，并进行了实验检验。
基于传统Canny算法中采用高斯滤波器对图像滤波平滑处理的效果有待改进，本论文引用了自适应中值滤波器，在使用Canny算法之前，对图像进行滤波，通过图3.4的结果显示，检测效果明显改善；而在取阈值时，在使用Otsu算法的基础上，采用顶帽算子对图像进行预处理以补偿图像，这样能取得更合适的阈值；图3.5和图3.6检测结果分别在滤波器、阈值两个方面进行了实验的比较，并与传统的Canny算法的检测结果进行了对比，显示出了本文所采用的改进算法的优越性。
关键词：边缘检测；Canny算子；自适应中值滤波器；Otsu算法；顶帽算子
I
基于matlab的图像边缘检测算法研究
Research on image edge detection algorithms based
on MATLAB
ABSTRACT
Image edge detection technology is the basis of digital image processing technology research, an important foundation for object recognition. Existing edge detection technique has some limitations in noise suppression threshold parameter selection, poor adaptive capacity needs to be further improved and enhanced.
This paper introduces the image edge detection significance and background of the subject; theoretical basis, in the second chapter a brief introduction of the traditional image edge detection algorithm, such as Roberts operator, Sobel operator, Prewitt operator, Laplacian operator LOG operator, recalling the classic edge detection algorithm, Canny algorithm described later as bedding. In Chapter 3, the basic principle of combining the Canny algorithm, the algorithm of the three criteria, the idea of the algorithm and the detecting step, the improved method of on Canny operator image smoothing filtering and thresholding algorithms, and experiments were carried out test.
Based on the traditional Canny algorithm using Gaussian filter smoothing effect for image filtering needs to be improved, the paper quoted the adaptive median filter, Canny algorithm, the image is filtered by the results of Figure 3.4 shows the detection results significantly improved; when the threshold value is taken using the Otsu algorithm based on the top-hat operator preprocessing on the image to compensate for the image, this can obtain a more appropriate threshold; Figure 3.5 and Figure 3.6 the detection result, respectively, in the filter threshold two aspects of the comparison of the experiment, and compared with the traditional the Canny algorithm of detection results show the superiority of the improved algorithm used in this paper.
Keywords: edge detection; Canny operator; adaptive median filter; Otsu algorithm; Top-Hat operator
II
