基于聚类和局部区域的彩色图像分割方法  

Segmentation method of color image based on cluster and local regions

作　　者：朱征宇[1,2] 王丽敏[1] 

机构地区：[1]重庆大学计算机学院,重庆400030 [2]重庆大学软件理论与技术重庆市重点实验室,重庆400030

出　　处：《计算机工程与设计》2015年第1期201-205,共5页Computer Engineering and Design

基　　金：科技部国家科技支撑计划重点基金项目(2011BAH25B041)

摘　　要：针对现有图像多阈值分割方法存在的分割不够准确、计算复杂度较高等问题,利用Otsu多阈值分割的思想,提出一种基于聚类和局部区域的彩色图像多阈值分割方法。获取彩色图像的H分量直方图;在综合考虑以H分量划分的颜色范围的前提下对直方图中的H分量进行聚类,依次选取各类中的最大值作为对应局部区域的边界;分别搜索局部区域内的最小H分量值,将其作为彩色图像的分割阈值。实验结果表明,与现有方法相比,该方法具有较高的分割准确性和更快的计算速度。Aiming at the problems of inaccuracy and higher computational complexity existing in the current multi-threshold seg- mentation methods of images, a multi-threshold segmentation method of color images based on the cluster and local regions was proposed. The thought of Otsu multi-threshold segmentation was applied. Firstly, the histogram of the color image based on the H component was acquired. Secondly, the H component in the histogram was clustered on the premise of considering the color range which was divided using the H component, and the maximum values of each class were chosen as region boundaries. Finally, the minimum values in the local regions were searched, and these minimum values were set as the image segmentation thresholds. The experimental results show that compared with current methods, this method has not only the higher segmentation accuracy, but the higher calculation speed.

关 键 词：图像分割 多阈值 聚类 最大值 最小值 局部区域 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]