Low-Cost Road Traffic State Estimation System Using Time-Spatial Image Processing

Abstract

Road traffic mobility can be described by its Level of Services (LoS). A major challenge in traffic state and LoS estimation is the limitation of observed traffic data. To derive the traffic state of a road network, a sensor network needs to be installed. Most stationary sensing techniques involve high investment in terms of the sensor installation, data communication and computational resources. This paper proposes a low-cost image processing system for road traffic state estimation using time-spatial image (TSI) processing. The TSI is an image processing technique for transforming a series of video images into a single image. Therefore, the TSI can reduce memory resources compared with the traditional methods. A camera can be exploited for traffic-state estimation through integration with TSI generating and processing modules. In addition, traffic state variables such as space-mean-speed, flow and density can be estimated. Empirical results are provided based on several experiments to show that TSI processing is a viable lowcost approach to traffic state estimation.