The retrieval method of multimedia database systems trends toward content-based way by automatic parsing. Although the automatic parsing of general video is very difficult, the parsing for a specific type of video can be achieved with the video structure which can be explicitly identified.
With the increase of the number of cars, we need a traffic management system which maximizes the efficiency of the road-usage. Therefore, we propose the parsing of traffic information video. The traffic information video is continuously captured from a camera which is established on a road. The traffic management system must have the function of filtering the information about traffic and cars by analyzing the traffic information video. Using the information, we can obtain the traffic information which will be stored in a database and we can control the violation of traffic regulation. So, a suitable parsing method for traffic information video is needed. The parser inputs the traffic volume to the semantic modeller and passes the information about the violation of traffic regulation to an indexing tool. We develped a method to parse the content of traffic information video using domain knowledge and implemented the modules which can detect the traffic volume, central lane violation, traffic lane violation, and signal light violation.