圆形直方图Otsu迭代阈值分割算法  

Otsu iterative thresholding segmentation algorithms on circular histogram

作　　者：王东 杨贵峰 范九伦 WANG Dong;YANG Guifeng;FAN Jiulun(School of Communications and Information Engineering,Xi an University of Posts and Telecommunications,Xi an 710121,China)

机构地区：[1]西安邮电大学通信与信息工程学院,陕西西安710121

出　　处：《西安邮电大学学报》2021年第3期53-60,共8页Journal of Xi’an University of Posts and Telecommunications

基　　金：国家自然科学基金项目(62071378,62071379,62171380)。

摘　　要：为了降低基于颜色空间H分量的圆形直方图阈值分割算法的复杂度,提高分割效果,基于高效圆形直方图Otsu法,将线性Otsu迭代算法扩展至圆上,提出圆形直方图Otsu法的线性均值迭代算法和线性均值嵌套式迭代算法。实验结果表明,所提的两个迭代算法在有效提取目标的同时降低了时间复杂度。In order to reduce the complexity of the circular histogram threshold segmentation algorithm based on the H component of the color space,and to improve the segmentation effect,the linear Otsu iterative algorithm is extended to a circle based on the efficient circular histogram Otsu method,and the linear mean iterative algorithm and the linear mean nested iterative algorithm of circular histogram Otsu method are proposed.Experimental results show that the proposed two iterative algorithms can effectively extract the target and reduce the time complexity.

关 键 词：圆形直方图 阈值分割 OTSU法 迭代算法 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]