首页 | 本学科首页   官方微博 | 高级检索  
     

红外热图像序列中基于人体模型的目标头部定位方法
引用本文:杨,裴继红,谢维信. 红外热图像序列中基于人体模型的目标头部定位方法[J]. 激光与红外, 2005, 35(2): 126-129
作者姓名:  裴继红  谢维信
作者单位:深圳大学信息工程学院,广东,深圳,518060;深圳大学现代教育技术与信息中心,广东,深圳,518060
摘    要:文章针对红外热图像序列,提出了一种多目标的头部定位方法。该方法检测目标轮廓的上凸点,根据上凸点构造粗略人体模型;利用热函数聚类方法确定投影直方图中峰的个数及位置。最后对上凸点和峰顶点进行综合判断以确定目标个数并定位。实验结果表明,该方法是有效可行的。

关 键 词:图像分割  目标定位  红外热图像
文章编号:1001-5078(2005)02-0126-04

Multi People Head Detection in Infrared Thermal Image Sequences Based on Human Shape Model
Yang Xuan,PEI Ji-hong,XIE Wei-xin. Multi People Head Detection in Infrared Thermal Image Sequences Based on Human Shape Model[J]. Laser & Infrared, 2005, 35(2): 126-129
Authors:Yang Xuan  PEI Ji-hong  XIE Wei-xin
Affiliation:YANG Xuan~1,PEI Ji-hong~2,XIE Wei-xin~1
Abstract:A novelmethod used to count and locate multi2moving objects in thermal sequences is p roposed in this paper. At first, convex points of objects boundary are detected. Thermal and position information of human shape areused to weight the coefficients of vertical p rojection histogram. Potential function clusteringmethod is used to determinate the number of apex in the histogram and to locate the position of apex. Conves points and apex of histogram are chosen to determinate the number and the position ofmoving object heads. Experimental results show that ourmethodis effective and feasible to count and locate the moving objects in thermal sequences.
Keywords:image segmentation  head detection  infrared thermal images
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《激光与红外》浏览原始摘要信息
点击此处可从《激光与红外》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号