基于自适应模糊聚类的图像分割算法答辩.docx

基于自适应模糊聚类分析的图像分割算法研究 摘要：图像分割是数字图像处理研究的重要领域，也是其研究的热点问题之。图像分割是把图像分割成各具意义的区域，它对于图像特征提取、图像分析与描述、图像识别等具有重要意义。目前，有许多种图像分割方法，本文提出了两种基于自适应模糊聚类的图像分割算法。聚类分析是一种无监督分类法，在缺少先验知识时，图像分割可以通过聚类分析完成。 本文讨论和分析了图像分割的研究现状，传统分割方法及其优缺点；针对聚类分析算法的特点进行了分析，着重研究了聚类算法及其在图像分割中的应用；在研究模糊C均值算法和均值漂移算法的基础上提出了两种自适应的图像分割算法。 （1）传统的模糊C均值聚类算法对初始值敏感，不良的初始值会导致算法的收敛速度过慢和收敛到局部极值。针对这些不足，本文将遗传算法引入到FCM算法中，提出了一种自适应初始值的FCM图像分割算法。利用遗传算法强大通用性，较好地解决了FCM算法对初始化敏感的问题，又能在一定程度上提高了图像分割的速度，获得了较好的分割结果。 （2）均值漂移算法是一种统计迭代的核密度估计方法，用于图像分割时不需要任何先验知识，近年来在图像分割领域得到广泛的应用。但是均值漂移算法中带宽的选择对算法性能有较大的影响，采用固定带宽，如果带宽选择过大，可能会合并某些极值点。因此本文提出了一种自适应带宽的均值漂移图像分割算法，获得了高质量的分割效果。 关键词：图像分割；聚类分析；自适应；模糊C均值；均值漂移 Abstract: Image segmentation is an important research field of digital image processing, and is one of the hot issues in the study of it. Image segmentation is just to segment an image into different sub-images with different characters. It is important significance for image feature extraction, image analysis and description, image recognition. At present, there are many kind of image segmentation method. This paper put forward two based on adaptive fuzzy clustering image segmentation algorithm. The clustering analysis is an unsupervised classification method. Clustering analysis is always carried out in the condition with no pre-known knowledge. In this paper, then clustering algorithms and image segmentation by clustering were researched in depth according to the characteristics of clustering. Through selection and improvement of clustering algorithms, use FCM algorithms and Mean Shift algorithm, two image segmentation algorithms based on auto-adaption clustering analysis are proposed. The main contents of the thesis includes: (1)Fuzzy C-means algorithm is one of the widely applied fuzzy algorithms at present. Based on the analysis of advantages and disadvantages of the self-adaptive FCM image segmentation algorithm, an improved self-adaptive fuzzy C-means clustering algorithm is proposed. First of all, the method adopts an initial algorithm to ass