Speed Calculation Of Traffic Accident Vehicle Based On Image Plane Movement

As an important part in the accident identified, the speed of vehicle can not only describethe nature of the accident, but also be conducive to reproduce the accident process. Based onthe location parameters, accident marks, scattered, etc, the traditional means of vehicleidentification are calculated by operator, and the identification process is complex, whichreduce the road capacity over a period of time. Therefore, the use of video speed accidentidentified, which is an intuitive, real-time, small impact on road traffic way, has a strongapplication value.By analyzing the calculation model and calculation steps of surveillance cameracalibration technique, moving target detection and tracking technology, this paper presents anew calibration method and moving object (vehicle) detection, tracking process. Then thepaper establishes the speed automated processing systems by VC++, and completes the test ofthe effectiveness of the system.First, this paper studies the surveillance camera calibration method. Through analyzingthe calculation process, the model and parameters of the linear and nonlinear methods ofsurveillance cameras, this paper presents a simple, stable calibration method based onaccumulation point and perspective projection relationship, conducting precision check.Second, this paper studies the moving target (vehicle) detection technology and trackingtechnology. Depending on the different model and calculation method, the paper analyzes theoptical flow method, the inter-frame difference method and background subtraction method.By comparing the advantages and disadvantages of the various techniques, the paper designsthe car fluid region detection model. Then the paper analyzes the movement of the center ofthe target vehicle after extraction by infinitesimal inertial motion model, designing the vehiclefluid region tracking model.Finally, this paper develops computing speed automated processing system. Based onVC++platform, this system includes: video image reading and extraction module, completingthe separation of the single-frame images from the video stream; space coordinatescalculation module based on the relationship of coordinates mapping; car fluid regionalpositioning module and tracking module, completing the accident vehicle speed computingsystem. At last, the system effectiveness test is completed.