Moving People Detection in Dynamic Scenes by Stereo Vision

• Published in: Intelligent Science and Intelligent Data Engineering pp. 506 - 512
• Main Authors: Zhuo, Tao, Zhang, Yanning, Yang, Tao, Zhang, Xiaoqiang
• Format: Book Chapter
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg
• Series: Lecture Notes in Computer Science
• Subjects: Histograms of Oriented Gradients; moving people detection; stereo vision
• ISBN: 3642319181; 9783642319181
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-642-31919-8_65

Robust people localization on a moving robot platform is an important and challenge research topic. In this paper, we present a novel moving people detection approach and system on a mobile robot platform. The proposed method mainly contains two parts: (1) A Histograms of Oriented Gradients (HOG) based detector is adopted to detect the human candidates in the dynamic scene; (2) Geometric constraints are computed to handle the ghost problem, including the depth information from a stereo camera and the external parameters of the camera calibration. To evaluate the proposed approach, a robust people detection system is developed on a moving robot platform. Extensive experiment results with challenge indoor and outdoor scenarios demonstrate the robustness and efficiency of our approach.