Multi-perspective Video Analysis of Persons and Vehicles for Enhanced Situational Awareness

This paper presents a multi-perspective vision-based analysis of people and vehicle activities for the enhancement of situational awareness. Multiple perspectives provide a useful invariant feature of object in image, i.e., the footage area on the ground. Moving objects are detected in image domain, and tracking results of the objects are represented in projection domain using planar homography. Spatio-temporal relationships between human and vehicle tracks are categorized to safe or unsafe situation depending on site context such as walkway and driveway locations. Semantic-level information of the situation is achieved with the anticipation of possible directions of near-future tracks using piecewise velocity history. Crowd density is estimated from the footage in homography plane. Experimental data show promising results. Our framework can be applied to broad range of situational awareness for emergency response, disaster prevention, human interactions in structured environments, and crowd movement analysis in wide-view areas.