The selection of a suitable feature is an important and indispensible issue in real-time surveillance systems employing a network of cameras to achieve rapid and accurate results. We have used the concept of color image histograms, as a simple but efficient method, to track people in different cameras in this research. Employing the image histogram for tracking may not be robust as it is sensitive to the environmental brightness changes and the human body shape and size. In this paper, we propose a novel method employing the Cumulative Brightness Function (CBF) to attenuate the effects of brightness changes in tracking objects moving in the camera networks. In the proposed method a fuzzy system is used to partition the field of view of each camera to cope with the size variations of people in different cameras. Then the body of each subject is divided into three portions, namely the head, the torso and the bottom section, with the prior knowledge about the human physiology, to have a better distinction between variance people appearance color. Finally, the histograms of the body sections with respect to the distance of subject from each camera are used for subject identification, reidentification and tracking. The results achieved by the proposed method in various environmental conditions show its robustness and effectiveness for tracking people in a network of cameras.
Rights and permissions | |
![]() |
This Journal is an open access Journal Licensed under the Creative Commons Attribution-NonCommercial 4.0 International License. (CC BY NC 4.0) |