Detection Algorithm for Real Surveillance Cameras Using Geographic Information

Abstract

Detection algorithm for pedestrians is proposed for the real surveillance system based on color similarity for dynamic color images under low illumination, where the proposed color similarity is defined by color change vectors in the L*a*b* color metric space and the time taken by pedestrians to pass between surveillance camera. It provides continuous detection results through surveillance cameras under lower luminance conditions in real surveillance system. Experimental results for dynamic image taken under low illumination in streets show that detected frames with the proposed algorithm increased by 20% compared to detection results without geographic information. The proposed algorithm is being considered for use in poor security areas in downtown Japan.