摘要图像分割是图像处理的基本内容之一，是把图像分割成各具特性（如灰度、颜色、纹理等特性）的子区域并提取感兴趣目标的技术和过程，是图像处理、图像分析和图像理解的基础。图像分割不仅一直都是图像技术研究的热点和焦点之一，而且在诸如计算机视觉、模式识别、医学图像处理及工业领域等实际中也得到了广泛的应用。65566

本论文针对亮度层次多的目标物体的分割问题，在介绍几种常用的阈值分割算法的基础上，首先实现了灰度直方图的绘制以及均衡化算法，为后文算法奠定基础；然后重点研究和实现了基于熵的单级和多级图像分割算法，并对两种算法的分割效果进行对比分析，得出多级分割比单级分割效果更好、适用范围更广泛的结论。

毕业论文关键词  图像分割 熵 多级

毕业设计说明书（论文）外文摘要

Title    Research and Implementation of an automatic thresholding image segmentation algorithm   

Abstract

Image segmentation is one of the basic contents of image processing, which is used to pide an image into sub-regions with specific characteristics (such as gray scale, color, and texture ) and extract targets of interest . It is the basis of image processing, image analysis and image understanding. Image segmentation not only has always been one research focus of image technology , but also has been widely used in practice such as computer vision, pattern recognition, medical image processing and industrial areas.

    In this thesis, the segmentation of targets with multi-level brightness is studied. On the basis of introducing several common used threshold segmentation algorithm,histogram calculation and equalization  are firstly achieved ,which lay the foundation for the later algorithm. Then a segmentation algorithm based on single-stage entropy and multi-stage entropy is realized,and the comparison and analysis of the experiment results are also carried out.At last, the conclusion shows that multi-stage segmentation has better performance and broader application .

Keywords: Image segmentation, Entropy, Multi-stage

目   次

1 引言 1

1.1 本课题的意义和背景 1

1.2 本论文的主要内容 1

2   阈值分割基础知识 1

2.1 阈值分割概述 1

2.2 典型的自动阈值分割算法 2

2.3 本章小结 7

3   基于单级熵与多级熵的阈值分割算法 7

3.1 数字图像基础 8

3.2 图像噪声的抑制 9

3.3 灰度图像的直方图及其均衡化 13

3.4 单级熵与多级熵分割算法实现 17

3.5 本章小结 26

结论 27

致谢 28

参考文献 29

1 引言

1.1 本课题的意义和背景

图像分割是图像处理的基本内容之一，是把图像分割成各具特性（如灰度、颜色、纹理等特性）的子区域并提取感兴趣目标的技术和过程，是图像处理、图像分析和图像理解的基础。[1] 显然，如果图像分割这一步处理得不好，那么，后续处理也会受影响。因此，图像分割不仅一直都是图像技术研究的热点和焦点之一，而且在诸如计算机视觉、模式识别、医学图像处理及工业领域等实际中也得到了广泛的应用。[1]