An Approach for Vehicle Warning Based on Video Information in Transportation Surveillance

The traffic safety monitoring technology becomes an active research field due to the development of the city´s transport system. In this paper, we design a warning method in traffic video monitoring through analyzing vehicle states. This method is based on image analysis and processing technology, extracting the information of the characteristics of the movement of vehicles through detecting and tracking moving vehicles in video image sequence. Then, with the characteristics of the movement of vehicles, we can approximate the trajectory of the vehicle by polynomial and predict the movement of vehicles. According to the velocity vectors of vehicles, we build models to determine the relationship between vehicles and warn the accident that may occur.