A Real-Time System for Monitoring Pedestrians

A Real-Time System for Monitoring Pedestrians
复制标题

DOI:
10.1109/acvmot.2005.10
复制
发表时间:
2005-01
期刊:
2005 Seventh IEEE Workshops on Applications of Computer Vision (WACV/MOTION'05) - Volume 1
影响因子:
--
通讯作者:
Huijing Zhao;R. Shibasaki
Huijing Zhao;R. Shibasaki
中科院分区:
其他
文献类型:
--
作者:
Huijing Zhao;R. Shibasaki

文献摘要

被引文献

相似文献

在本研究中,我们提出了一种新颖的行人监控系统,可以在大型开放区域(如商场、展览馆、火车站等)中跟踪行人的轨迹。利用单排激光测距扫描仪网络,将其设置在地板上,在高于地板约15厘米的高度进行水平扫描。定义了行人在地面水平截面上的行走模型,并开发了一种跟踪算法来检测和跟踪有节奏的摆动脚。另一方面,将摄像机与激光扫描仪网络相结合,对激光扫描仪测量覆盖范围内的某些点进行聚焦,获取行人轨迹的图像数据。在一个大型展览中进行了实时监测参观者流量的实验,该展览在三天的时间里总共有3万名参观者。四台LMS200 (SICK)和一台摄像机位于一个过境点附近,占地面积分别约为30 m × 30 m和5 m × 6 m。参观者和附近的摄像机的视野被实时跟踪。他们的轨迹以及脚上的激光点被投影到空中的视频图像上。另一方面,通过分析每个游客在离线模式下的运动行为,考察过境点附近的游客流量及其随时间的变化
In this research, we propose a novel system of monitoring pedestrians and tracking their trajectories in a wide and open area, such as shopping mall, exhibition hall, railway station and so on. A network of single row laser range scanners is exploited, which are set on the floor doing horizontal scan at an elevation of about 15cm above the floor. A pedestrian's walking model on the horizontal cross section of ground level is defined and a tracking algorithm is developed to detect and trace the rhythmic swing feet. On the other hand, video cameras are integrated with the network of laser scanners, which focus on some spots inside laser scanner's measurement coverage, and get image data for pedestrian's trajectories. An experiment for real-timely monitoring visitor's flow was conducted in a large exhibition, which had totally 30,000 visitors during three exhibition days. Four LMS200 (by SICK) and a video camera were located near a crossing point, covering a floor area of about 30 m times 30 m and 5 m times 6 m respectively. Visitors inside and nearby video camera's vision field were tracked in a real-time mode. Their trajectories as well as the laser points on their feet were back projected onto the video images on the air. On the other hand, visitor's flow nearby the crossing point and its change along with time is examined by analyzing the motion behavior of each visitor in an off-line mode