Group-based Multiple People Tracking in Perception Sensor Network

Group-based Multiple People Tracking in Perception Sensor Network
Le Anh Vu안기진최종석
human tracking; sensor fusion; robot operation system; group tracking; feature matching
Issue Date
Identification of human information such as position and name plays an importance role to effectively enhance interaction tasks between a robot and human. In the multiple human tracking scenarios, tracking a single human is a difficult task for a robot when people present close each other. In this paper a novel method is proposed to help a robot to keep tracking people even in the group situation. Specifically, a perception sensor network (PSN) system including multiple Kinects and PTZ cameras tracks multiple people based on grouping and ungrouping algorithm. When a group of multiple people staying close together is formed, individual features such as 3d human location, height, color and SURF feature points of human region of interest (ROI), which are identical criterions retrieving from both depth and color images of group member, are then stored and updated into the group database. Based on the distance between a group center and each member in the group, PSN system decides whether to keep this member in the group or to ungroup it from the group then reassign the right name identification from group database by minimizing multiple criteria. The experimental results demonstrate the proposed method outperforming conventional methods such as Kalman filter-based human tracking and 3d point-based human tracking.
Appears in Collections:
KIST Publication > Conference Paper
Files in This Item:
There are no files associated with this item.
RIS (EndNote)
XLS (Excel)


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.