Automatic Incident Detection Based On Vedio

Along with the fast development of the social economy, Intelligent Transportation System (ITS) becomes more and more important to our daily life. The Traffic Automatic Incident Detection System (TAIDS), which is one of the key parts of ITS, is able to detect and deal with the traffic incidents on the roads, and to decrese the death rate and weath loss. Furthmore, TAIDS can help to avoid subsequent incidents, save energy sources, reduce pollution and so on. As the core of TAIDS, Traffic Incident Detection Algorithm (AID) is a crucial problem and worthy of study.The dissertation stems from subproject "the study on Automatic Incident Detection in intersections based on vedio" of the item "the study on Automatic Incident Detection technology in urban intersections based on multi-source integration", which is funded by National "863" high-tech research and development project (2007AA11Z224). According to the characters of traffic flow and the fundamental requirements of traffic incident detection, the techniques and theories relevant to traffic incident detection are systematically investigated in this thesis.Firstly, it is reviewed that the classification of traffic incidents. Moreover, kinds of traffice parameters acquisition technologies are compared and discussed. And then an overview of AID algorithm research history and current situation are pointed out. Focusing on intersection, a particular traffic scence, an incident detection approach based on analysis the stop reason of the terminal car is presented, in accordance with the traffic flow characteristics. Meantime, the traffic incident detection system in intersection basen on vedio is designed and completed partly. On the basis of vehicle detection and tracking, traffic parameters extraction methods are found for reserving sufficient data. This paper proposed an new AID algorithm called SVM-AID after. Using actual data I-880 database, the experiment researches the applicability of the support vector machines (SVM) in actual traffic AID systems and how the parameters worked in SVM-AID. The results showed that SVM-AID can perform well in detection rate and mean time to detection when the parameters are appropriate. In conclusion, the dissertation concludes the whole work, analyses the findings and achievements, and brings forward further research directions and projects.