Title :
Online multi-target tracking using non-linear motion estimation
Author :
Sijong Kim ; Myung Jin Chung
Author_Institution :
Dept. of Electr. Eng., KAIST, Daejeon, South Korea
Abstract :
This paper deals with an online multi-target tracking method applicable to real world applications such as surveillance systems, traffic monitoring, and driving assistance systems. The proposed method consists of three stages, non-linear motion estimation in continuous time domain, hypothesis generation, data association. We tested the proposed method on publicly available video sequences and analyzed the results of the proposed method.
Keywords :
heuristic programming; image sequences; motion estimation; object tracking; sensor fusion; target tracking; video signal processing; continuous time domain; data association; hypothesis generation; nonlinear motion estimation; online multitarget tracking; video sequence; Computational modeling; Logic gates; Motion estimation; Target tracking; Time-domain analysis; Trajectory; data association; non-linear motion; online multi-target tracking;
Conference_Titel :
Ubiquitous Robots and Ambient Intelligence (URAI), 2014 11th International Conference on
DOI :
10.1109/URAI.2014.7057490
