| 多视角多行人目标检测、定位与对应算法 |
| |
| 引用本文： | 徐剑,丁晓青,王生进,吴佑寿.多视角多行人目标检测、定位与对应算法[J].清华大学学报(自然科学版),2009(8). |
| |
| 作者姓名： | 徐剑 丁晓青 王生进 吴佑寿 |
| |
| 作者单位： | 清华大学电子工程系; |
| |
| 基金项目： | 国家自然科学基金资助项目(60472002) |
| |
| 摘 要： | 为解决多视角视频监控中多行人情况下的遮挡问题,引入三维重建的思想,提出一种基于空间场的多视角多行人检测、定位和对应算法。该算法首先提取各视角运动前景,融合各视角二值前景图像,利用空间场实施重建,然后根据信息融合结果在空间中检测目标并定位,并由空间检测和定位结果确定各视角中目标的对应关系。在3dsM ax合成数据和实际采集数据上进行的实验测试表明,该算法对图像遮挡的处理能力强,计算复杂度低,基本满足实时要求。
|
| 关 键 词： | 多视角 空间场 遮挡 检测 对应 定位 |
Detection,location and labeling under a multi-moving-person,multi-view set |
| |
| XU Jian,DING Xiaoqing,WANG Shengjin,WU Youshou.Detection,location and labeling under a multi-moving-person,multi-view set[J].Journal of Tsinghua University(Science and Technology),2009(8). |
| |
| Authors: | XU Jian DING Xiaoqing WANG Shengjin WU Youshou |
| |
| Institution: | Department of Electronic Engineering;Tsinghua University;Beijing 100084;China |
| |
| Abstract: | A multiple people detection,location and labeling algorithm in a multi-moving-person,multi-view set was developed to solve the problems of occlusion.The algorithm bases on a 3-D reconstruction algorithm with a space field model based framework.It extracts moving foregrounds from respective views and fuses them to reconstruct 3-D space field firstly,and then analyses of the space field results lead to detection and location of the humans in 3-D space which are then labelled in all views in the 2-D image spac... |
| |
| Keywords: | multi-view space field model occlusion detection correspondence location |
| 本文献已被 CNKI 等数据库收录！ |
|
