Effective Estimation of Histogram Differenced Value using Multiple Contiguous Virtual Layer (MCVL) for ITS Applications

For successful operation, any video-based Intelligent Transportation Systems (ITS) application requires real-time road traffic information or characteristics such as speed, density, average delay, categorization, and so on. This paper proposes the Multiple Contiguous Virtual Layer (MCVL), a robust and unique vehicle identification framework that estimates any macroscopic traffic characteristics using computer vision algorithms on traffic video. This work mainly focuses on estimating a new parameter known as Histogram Differenced Value (HDV) for MCVL, which uses spatial color information to reveal substantial differences in traffic condition. Several benchmark traffic video datasets are used to test the performance and accuracy of estimations utilizing the proposed framework, with the results being discussed. The results indicate that using the proposed HDV parameter, the accuracy of the vehicle recognition process is improved with the combination of lowered computing cost of MCVL.