Real-Time Non-Motor Vehicle Violation Detection in Traffic Scenes

The blockchain technology in the traffic scene is developing rapidly recently. In this area, distributed system based technology like blockchain has been standing out with its decentralization, data-innovation and data transparency. We propose a real-time non-motor vehicle violation detecting approach, which can detect and collect violation information of non-motor vehicle and has been integrated into the distributed smart traffic management system. We propose a luminance balancing in the pre-processing steps to address the problem of brightness change in the background. We use the refined codebook algorithm to detect the motion part, and then propose a violation judging criteria based on feature matching to obtain the violation information. Experiments and application have shown the effectiveness and efficiency of our approach.