Detection and tracking an object in omni-directional images using particle filter.

Autor: Katayama, Koutaro, Shibata, Keiji, Horita, Yuukou
Zdroj: 2012 Proceedings of SICE Annual Conference (SICE); 1/ 1/2012, p1040-1043, 4p
Abstrakt: It is predicted that demand for surveillance cameras become higher, and accordingly, object detection early and tracking object accurately are important. This paper proposes detection and tracking method for the omni-directional images taken by a camera with PAL. These method use particles filter for detection and tracking. In detection, the method uses that an object appears from the outer periphery of the imaging area. In tracking, the other method uses the estimation of the object area conforming to the shape the omni-directional images. As a result, the detection method has got to detect an object earlier, and the tracking method has tracked the object with high accuracy. [ABSTRACT FROM PUBLISHER]
Databáze: Complementary Index