Multi-feature Detection for Pedestrian Tracking in Traffic Surveillance
Xuan hang Wang, Song Huansheng, Cui Hua · 2016
Pedestrian tracking is an active research area to improve traffic safety for intelligent video surveillance. This paper proposes an efficient method to automatically detect and track far-away pedestrians in surveillance video using the motion feature extraction and analysis. Firstly, pedestrian features of each frame are extracted by object segmentation, recognition and feature extraction. Then, the similar features in current frame image of all candidate objects are matched by the characteristic information of pedestrians in the previous frame which is considered as a template. Finally, pedestrian trajectory analysis algorithms are used on the track trajectories and the motion information can be attained, which can realize the early classification warning of pedestrian events. Experimental results in practical surveillance demonstrate that this method shorten the processing time of matching pedestrians and improve the reliability and real-time ability of pedestrian tracking.