Pedestrian Tracking in Infrared Image Sequence Based on Spatio-temporal Slice Trajectory Analysis

The research of pedestrian tracking algorithm in infrared image sequence is a curial part in video surveillance. But due to the special characteristics of infrared image, such as low contrast, fuzzy edge and unknown noises interference, the study of infrared pedestrian tracking algorithm becomes a great challenge. Spatio-temporal slice method is an effective analysis method considering both space and time. In view of the existing tracking methods based on spatiotemporal slice only considering horizontal slice analysis and usually with large amount of calculation, this paper proposes a novel tracking algorithm based on spatio-temporal slice trajectory analysis. The proposed algorithm uses multi-layer horizontal and vertical slices to obtain the complete trajectories in order to determine the information of target boundary and position, and then tracks the target in each frame of the infrared image sequence. The experimental results show that the algorithm has relatively high tracking accuracy at a fast computing speed. Moreover, it performs effectively in different infrared image sequences with various motion modes by single pedestrian from OTCBVS/05 Terravic Motion IR Database, namely, the algorithm has good robustness to some extent.