摘要图像分割是把图像分割成若干个各具特性的区域并提取出感兴趣的目标的技术，这些特性可以是像素的灰度、颜色、纹理等，提取的目标可以是单个区域，也可以是多个区域。图像分割技术在图像工程中的地位是不可替代的，它是图像处理和图像分析之间的关键步骤，也是进行更高级的技术的基础。同时，图像分割也是一个经典难题，在过去的几十年里,成百上千种的分割算法已被呈现在各种文献中，但是缺乏一个统一的理论体系，到目前为止还不存在一种通用的、能使各种类型的图像达到最优分割质量的图像分割方法，也不存在一种被广泛接受的方法，能够对不同的分割结果给出与主观感觉一致的评价。9361
本文对图像分割算法进行了较深入的研究。论文的主要工作如下：
(1)收集整理了国内外相关资料，将图像分割方法分为三类：基于边缘的分割算法，基于区域的分割算法，基于特定理论的分割算法。分别介绍了这三类算法中的常用算法。
(2)采用MATLAB 仿真了所有分割过程，得到了比较理想的分割结果，并通过比较，分析了各个算法的优点和不足之处,以及适用于何种图像。
关键词 图像分割 边缘检测 区域分割 分水岭
毕业设计说明书（论文）外文摘要
Title Study on image segmentation
Abstract
Image segmentation partitions an image into different regions with some specific properties and extracts the target we are interested in. These properties can be the pixel gray scale, color, texture and so on. The target can be a single region or multiple areas. The technique of image segmentation is very important in the project of image, it is a critical step between image processing and image analysis. It is also the basic of more advanced technology. However, image segmentation is a classic problem. During the past few decades, thousands of the segmentation algorithms have been presented in the literature, but we still lack a unified theoretical system. A common algorithm which can make various types of images achieve the optimal segmentation quality hasn’t been set up.
This paper studies the classic segmentation methods:
(1) Based hundreds of algorithms presented in the literature, this paper classify image segmentation algorithms into three categories: methods based edge, methods based region, and methods based special theories. Some common and classic algorithms belonging to the categories respectively are introduced.
(2) Segmentation process is simulated and the results have shown perfect. Advantages and disadvantages of each algorithm are discussed at the end of the paper, and to apply to each image.
Keywords Image segment Edge detection Region segment Watershed
1 引言 1
2 图像分割算法的发展和现状 3
3 图像预处理 4
3.1 均值滤波 5
3.2 中值滤波 7
4 基于边缘的分割算法 9
4.1 Robert 算子 10
4.2 Prewitt 算子和Sobel算子 12
4.3 canny 算子 14
4.4 Log 算子 15
4.5 各种边缘检测算子比较 15
4.6 边缘连接 16
5 基于区域的分割算法 17
5.1 阈值分割 17
5.2 区域分割 25
6 基于特定理论的分割算法 29
6.1 分水岭分割的基本思想和常用方法 29
6.2 分水岭分割实验及改进 33
结论 35
致谢 36
参考文献 37
附录 39 MATLAB 图像分割算法的研究+分水岭算法:http://www.751com.cn/zidonghua/lunwen_8066.html