3D-sensing Distributed Embedded System for People Tracking and Counting

Abstract

The present study focuses on the development of an embedded smart camera network dedicated to track and count people in public spaces. In the network, each node is capable of sensing, tracking and counting people while communicating with the adjacent nodes of the network. Each node typically uses a 3D-sensing camera positioned in a downward-view but the designed framework can accept other configurations. We present an estimation method for the relative position and orientation of the depth cameras. This system performs background modeling during the calibration process, using a fast and lightweight segmentation algorithm.