一种基于改进层次聚类的Hough空间峰值检测方法  

Peak Detection Method of Hough Space Based on Improving Hierarchical Clustering Method

作　　者：陈东[1] 林建粦[1] 马德宝[1] 

机构地区：[1]信息工程大学信息工程学院,河南郑州450002

出　　处：《信息工程大学学报》2010年第2期210-213,共4页Journal of Information Engineering University

基　　金：科研基金资助项目

摘　　要：运用Hough变换对序列图像中的弱小点目标运动轨迹进行直线检测时,常规的Hough变换参数空间峰值检测方法难以区分相近的目标运动轨迹。为满足多目标有效跟踪的应用需求,文章根据目标Hough变换参数空间的狭长型特点,结合改进的层次聚类方法,给出一种基于层次聚类的Hough变换参数空间峰值检测方法。通过实验分析,验证了文章所提方法的有效性。In making line detection of moving track of dim-point targets in image sequences with the Hough transform,the normal peak detection method of parameter space of Hough transform is hardly able to distinguish the moving track of targets that are different but adjacent.In order to satisfy the application demand to tail the multiple targets efficiently,this paper presents a mend peak detection method of parameter space of Hough transform based on an improved hierarchical clustering method,which is also given for the long and narrow shape characteristic of Hough transform parameter space.Finally,an experiment analysis is made to testify the validity of the given method.

关 键 词：HOUGH变换 峰值检测 层次聚类方法 遍历 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]