(基于小波变换医学图像边缘检测与分析).doc

基于小波变换的医学图像边缘检测与分析 摘要 医学图像在形成、传输或变换过程中,由于受多种因素的影响,经常会使图像模糊不清,从而影响了医学图像在临床诊疗过程中的应用医学图像增强是医学影像处理研究的一个重要课题,它的目的就是提高图像的可懂度,使图像更适合人眼观察和机器识别 本文首先介绍了图像增强方面的基础知识，论述了常用的几种边缘检测算子的原理，主要是传统的边缘检测算子和线性滤波边缘检测算子。传统的边缘检测算子是基于一阶导数的边缘检测算子，代表性的有Roberts算子、Sobel算子、Prewitt算子等。线性滤波边缘检测算子其中最有代表性的是LOG算子和Canny算子。本文选取“CT.dcm”图像进行边缘检测,并分析比较它们的检测结果。分析算法的优缺点并提出了改善方法,即基于小波变换模极大值多尺度边缘检测方法，最后经比较发现其边缘检测效果好，在时域和频域同时具有良好的局部化特性，且其增加的计算复杂度小。 关键词：小波变换 边缘检测 图像处理 Abstract: When forming or transforming Medical images,due the influence of various factors,the images usually are fuzzy.so it influence the effects of medcial images in clinical application.Medical image enhancement is an important research topic of medical images processing,its purpose is to improve the intelligibility of the images,makes the images more suitable for human observation and machine recongnition. This paper firstly introduces the basic knowledge of image enhancement, discusses the theory of several commonly used edge detection operators,it is mainly the traditional edge detection operators and linear filter edge detection operators. The traditional edge detection operators are based on the first order derivative edge detection operators, which represented by the Roberts operator, Sobel operator, Prewitt operator. To the Linear filter edge detection operator,one of the most representative is the LOG operator and Canny operator. This paper selects the CT.dcm image,analysis and comparison of their results. Finally, analysis of the existing algorithms and put forward the improvement method, which is based on modulus maxima of wavelet transform multi-scale edge detection method, finally through the comparison of the edge detection effect ,we can see the method of wavelet is good, both in time and frequency domains has good localization characteristics, and its computation complexity is little than others. Key words: wavelet transform edge detection image process 目录 摘要 1 1 绪论 3 1.1引言 3 1.2研究现状与发展趋势 4 1.3研究内容 5 1.4论文结构 5 2 边缘检测算子及小