内容发布更新时间 : 2024/12/24 4:29:10星期一 下面是文章的全部内容请认真阅读。
毕业设计(论文)
图像分割算法研究与实现
学 院
专 业:
学 号: 姓 名:
职 称: 指导老师:
中国 二○一二年五月
讲师
诚信承诺书
本人郑重承诺:我所呈交的毕业论文《图像分割算法研究与实现》是在指导教师的指导下,独立开展研究取得的成果,文中引用他人的观点和材料,均在文后按顺序列出其参考文献,论文使用的数据真实可靠。
承诺人签名:
日期: 年 月 日
II
图像分割算法研究与实现
摘 要
数字图像目标分割与提取是数字图像处理和计算机视觉领域中一个备受关注的研究分支,其中最主要的是运用MATLAB对图像进行仿真分割,并用各个方法进行分析、对比并得出结论。本文主要介绍了图像分割的基本知识,从原理和应用效果上对经典的图像分割方法如边缘检测、阈值分割技术和区域生长等进行了分析。在边缘检测时对梯度算法中的Sobel算子、Prewitt算子、LoG(Laplacian-Gauss)算子、Canny算子的分割原理逐一介绍并比较各种算子的分割效果。而阈值分割技术的关键在于阈值的确定,只有阈值确定好了才能有效的划分物体与背景,本文着重实现基于迭代法的全局阈值及基于Otsu最大类间方差算法的自适应阈值。此外还介绍了区域增长法,它的基本思想是将具有相似性质的像素集合起来构成新区域。与此同时本文还分析了图像分割技术研究的方向。
关键词:图像分割 MATLAB 边缘检测 区域生成 阈值分割
III
Research of Image Segmentation Algorithm
ABSTRACT
Digital Image Object Segmentation and Extraction is a major concern in the field of digital image processing and computer vision research branch, which the most important is the use of MATLAB for image segmentation and simulation, using each method to carry on the analysis, comparison and conclusion. This paper mainly introduces the basic knowledge of image segmentation, based on the principle and the application effect to the classic image segmentation methods such as edge detection, threshold segmentation and region growing is analyzed. In the edge detection of gradient algorithm in the Sobel operator, Prewitt operator, Log operator, Canny operator segmentation principles introduced and comparison of various operators segmentation. While the threshold segmentation technology is the key to determine a threshold , only a good threshold can effectively divide object and background, this paper focuses on the implementation of the global threshold based on iterative algorithm and based on Otsu adaptive threshold algorithm. It also introduces the regional growth method, its basic idea is to have similar properties to the pixel together constitute a new area. At the same time the paper also analyzes the research direction of image segmentation technology.
Key words:Image segmentation MATLAB Edge detection Regional generation Threshold segmentation
IV