Low complexity techniques for robust real-time traffic incident detection

Abstract

Traffic congestion is one of the leading reasons for the development of intelligent transportation systems(ITS). Traffic incidents are the second biggest cause of traffic congestions after infrastructural bottlenecks. Real-time traffic incident detection for timely clearing of roads is required to ensure smooth traffic flow. Apart from the real-time performance, scalable solutions which can monitor wide areas in a cost-effective manner are required. In this paper, robust, lean and real-time stationary foreground object detection technique to detect traffic incidents has been presented. We use block-based analysis in contrast to the conventional pixel-based analysis to lower the computational complexity of the proposed technique and achieve real-time performance. Experimental evaluations on widely used datasets demonstrate that the proposed method can achieve comparable accuracy to the existing state-of-the-art techniques. The real-time performance of the proposed system has also been demonstrated by implementing it on a low-cost embedded platform, Odroid XU-4, that still achieves a frame rate of 40 frames/second, thereby enabling real-time detection of traffic incidents.