Font Size: a A A

Research On The System Of Traffic Safety Warning Based On Video

Posted on:2015-05-05Degree:MasterType:Thesis
Country:ChinaCandidate:X LiFull Text:PDF
GTID:2272330422985382Subject:Traffic Information Engineering & Control
Abstract/Summary:PDF Full Text Request
In traffic life, people, vehicles, road and environment are four important factors whichaffect the traffic safety, so the establishment of an effective traffic safety early warning systemis an research key in intelligent transportation field, to accurately detect the factors above andwarning timely,so as to guide drivers to drive safely and traffic management department tomanage effectively. In this paper, combining the technology of machine visual perception,pattern recognition, information fusion etc., a traffic early warning system based on trafficvideo detection is proposed to achieve the accurate detection of abnormal events which affecttraffic safety and the extraction of mainly traffic parameters, and promptly issued a safetywarning for the driver.The proposed system is composed with the traffic incident detection system andvehicle-road communication system. In the traffic incident detection system, processing thetraffic video images using image processing techniques to detect five kinds of incidents,including the retrograde, pedestrians, falling objects, parking and traffic congestion detection,at the same time parameters of such traffic conditions as traffic flow, speed are extracted;using the information fusion theory and decision making mechanism based on D-S evidencetheory to realize the collection, fusion and interactive processing of the dynamic traffic stateinformation. In the vehicle-road communication system, using the RFID roadside reader toachieve the vehicle authentication and registration, complete the networking, and report theone-time traffic issue; using ZigBee wireless communication device as the DSRCcommunication module to realize short distance wireless transmission of information betweenthe server and the terminal, and transfer the test results and vehicle information real time ingraphic form through the vehicle terminal to the driver.The proposed system has been built in the laboratory environment and the relatedfunctions and performance testing has been realized. Test results show that the proposedsolution and related algorithms can effectively and timely detect various events in trafficscene, and have the ability to correctly distinguish the different events, thereby to achieve thetraffic safety warning function.
Keywords/Search Tags:traffic safety warning, traffic incident detection, information fusion, vehicle-road communication, RFID, DSRC, vehicle terminal
PDF Full Text Request
Related items