双集合FCM和改进Otsu算法的灰度图像阈值分割  

Threshold Segmentation of Intensity Image Based on FCM Algorithm Owning Double Sets and Improved Otsu Algorithm

作　　者：李群力[1] 杜亚勤[1] 

机构地区：[1]西安工业大学计算机科学与工程学院,西安710032

出　　处：《西安工业大学学报》2012年第5期394-399,共6页Journal of Xi’an Technological University

摘　　要：针对FCM算法和Otsu算法的不足,根据灰度图像的特点,提出了基于双集合FCM和改进Otsu算法的阈值分割算法,该算法从两个方面对FCM算法和Otsu算法进行改进.第一、采用双集合的方法,减小FCM算法和Otsu算法的时间复杂度;第二、适当移动分割阈值提高Otsu算法的分割效果.从数学分析和实验可知,文中算法的时间花费和分割效果均优于FCM算法和Otsu算法.According to the disadvantages of FCM and Otsu algorithms and The characteristic of intensity image, an algorithm for threshold segmentation was presented on the basis of the FCM algorithm with double-sets and the improved Otsu algorithm. Both algorithms were improved in two aspects. First, double-sets is used to reduce the time complexity of both algorithms. 5;econd, the threshold of intensity image is adjusted properly to improve the effect of threshold segmentation of intensity image. The mathematical analysis and experiments show that the time efficiency and the effect of threshold seqmentation of the new algorithm are greater than both FCM and Otsu algorithms.

关 键 词：图像分割 模糊聚类 FCM算法 双集合 OTSU算法 阈值 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]