融合加窗色调直方图的彩色图像FCM分割算法  

Segmentation FCM algorithm combined hue histogram added window

作　　者：柳佳雯[1] 梁光明[2] 刘任任[1] 谢雅婷[1] 丁建文 

机构地区：[1]湘潭大学信息工程学院,湖南湘潭411105 [2]国防科技大学电子科学与工程学院,长沙410073 [3]长沙爱威科技,长沙410013

出　　处：《计算机工程与应用》2013年第3期230-233,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.60673193)

摘　　要：针对传统FCM算法计算量大、对噪声敏感的缺点,提出了一种融合加窗色调直方图的快速FCM算法(MHFCM)。对彩色图像的色调直方图加窗得到背景色调,对去除背景信息后的有效信息进行FCM聚类分割。实验结果表明,MHFCM算法提高了图像的分割速度,并且克服了图像中背景噪声对分割效果的影响,使分割更加准确。This paper proposes a fast FCM algorithm（MHFCM）combined hue histogram added window, to direct at conventional FCM algorithm which is characterized by the great amount of calculation, inferior resistance to noise. Hue histogram with added window is used for getting background hue from color images. Efficacious information without background information is used for clustering on FCM. Experiment results show that the MHFCM algorithm enhances segmentation speed and over- comes the effects of background noise on segmentation results to segment more precisely.

关 键 词：彩色图像分割 加窗色调直方图 模糊C均值聚类算法(FCM) 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]