灰度图像边缘特征分析及其提取毕业论文.doc

PAGE 本 科 毕 业 论 文 论文题目： 灰度图像边缘特征分析及提取 摘 要图像的边缘作为图像的一种基本特征，经常被应用到较高层次的图像应用中去。它在图像识别，图像分割，图像增强以及图像压缩等的领域中有较为广泛的应用，也是它们的基础。图像边缘是图像最基本的特征之一，往往携带着一幅图像的大部分信息。而边缘存在于图像的不规则结构和不平稳现象中，也即存在于信号的突变点处，这些点给出了图像轮廓的位置，这些轮廓常常是我们在图像边缘检测时所需要的非常重要的一些特征条件，这就需要我们对一幅图像检测并提取出它的边缘。而边缘检测算法则是图像边缘检测问题中经典技术难题之一，它的解决对于我们进行高层次的特征描述、识别和理解等有着重大的影响；又由于边缘检测在许多方面都有着非常重要的使用价值，所以人们一直在致力于研究和解决如何构造出具有良好性质及好的效果的边缘检测算子的问题。经典的边界提取技术大都基于微分运算。首先通过平滑来滤除图像中的噪声，然后进行一阶微分或二阶微分运算，求得梯度最大值或二阶导数的过零点，最后选取适当的阈值来提取边界。本文主要介绍几种经典的边缘提取算法，选取两种用MATLAB语言编程实现，对提取结果进行比较和分析。关键词: 边缘特征 图像边缘检测 算法 Abstract The edge of the image as an image of a basic feature, is often applied to a higher level of image application. It is in image recognition, image segmentation, image enhancement and image compression and other fields in a relatively wide range of applications, but also their basis. Image edge is one of the most basic features, often carries a lot of image information. While the edge in image of the irregular structure and unstable phenomenon, also is present in the signal mutation points, these points are given the image contour position, these contours are often in image edge detection are needed when the most important characters of condition, which requires us to image detection and extract the edges of it. Edge detection algorithm of image edge detection problem is a classic one of technical difficult problem, its solution for our high level description, recognition and understanding, have significant effect; and the edge detection in many aspects of a very important use value, so people have been committed to research and solve how to structure a good character and good effect of edge detection operator problems. Classic boundary extraction techniques are based on the differential operation. Firstly, through smoothing to filter the noises in the image, and then a differential or two order differential operators, obtained gradient maximum or two derivative of