Esearch On Detection Of Abandoning Items From Vechiles

In recent years, the intelligent video surveillance is playing an important role in transportation systems. The traffic event detection, as a critical application of intelligent video surveillance, is often used in the the traffic surveillance and suppressing the real-time traffic violations. This paper focuses on the event detection method for abandoning behaviors of vehicles by using the stationary camera. The method we proposed contains the detection of abandoned objects and the classification of abandoning subjects. Moreover, our method is successfully implemented in Motor Vehicle Intelligent Surveillance System.In this paper, the main works includes the following aspects:First, we define the event detection as follows:detecting the abandoned object, finding the key frames of the abandoning event, identifying and classifying the abandoning subject. Second, we propose a multi-layer method to detect the static abandoned object. Third, according to the characteristics of abandoning subjects in the scene of traffic (people and vehicles), we propose a support vector machine based method to classify the subjects. Finally, we design and implement the module of abandoning event detection, and successfully integrate it into the intelligent vehicle surveillance system. Experimental results show that the event detection module can implement real-time alarm, and the rate of accuracy is above 90%.