Research On Measurement System Of Vehicle Curve Speed Based On Video

Calculating the vehicle speed based on video accurately is of great significance for the identification of traffic accidents.With the popularity of traffic monitoring equipment,calculating the speed of accident vehicles by video has become a convenient technology.Vehicle curve speed detection is the difficulty in video detection technology,so the research in this paper has important use value.Firstly,through the research of video-based vehicle curve speed detection technology at home and abroad,this thesis finds that the curve speed error factors through the analysis of the trajectory of the feature points.Then Solving the vehicle heading angle and yaw rate based on the principle of perspective transformation in computer vision.The error caused by the inner wheel difference has been corrected by the vehicle heading angle and the yaw rate,and finally the vehicle curve speed calculation model has been established.Secondly,in order to verify the accuracy of the model,a vehicle test has been conducted at the automobile test site of Chang'an University.The vehicle has turning at different speeds and accelerations in test.The results show that the average relative error of the model is 3.1% and the maximum relative error is 6.6%.Finally,based on Python,the vehicle curve travel speed measurement system has been developed.The system comprises: a video processing module,a feature point selection and tracking module,a parameter input module,a curve speed calculation module and a result output module,realize the automatic measurement of the curve-based vehicle speed based on video.