Research And Implementation Of Vehicle Speed Measurement Based On Binocular Vision

At present,the car ownership in our country is increasing and the development of traffic management system is relatively backward.Traffic accidents have become one of the major public safety hazards in our country.The speeding of vehicles is one of the most important reasons,so the speed measurement has become a problem to be solved urgently.The traditional vehicle speed measurement is mainly based on radar speed,ground coil speed and laser speed.They have strict requirements for installation and high cost in maintenance,and the speed measurement system based on binocular vision is simple and easy to install.In the maintenance of the system,the precision of speed measurement can be greatly improved by upgrading the algorithm software,and the expensive hardware equipment is not needed to upgrade and upgrade.After studying the current method of binocular vision measurement,this paper adopts a combination velocity measurement algorithm,and designs a measuring system that can monitor the speed of the vehicle in real time.The main contents of this paper are as follows:On the basis of the camera imaging model,the current binocular vision velocity measurement algorithm is analyzed and studied.The hybrid Gauss background modeling algorithm is used to separate the moving vehicle targets,the SIFT corner detection algorithm is used to extract the stable feature points of the moving vehicles,the Euclidean distance matching algorithm is used to find the homonym points in the feature points,and the camera calibration is used.The equation is used to calculate the distance between vehicle and camera,and finally the vehicle speed is obtained.The working principle of the peripheral circuit of ARM+DSP dual core video processor platform and the multimedia framework provided by TI company are emphatically analyzed,and the cross compiling environment for the development of the system is built.The function of modifying the parameters of the hybrid Gauss background modeling algorithm is designed on the ARM processor to adapt to the scenes with different complexity,and the speed measurement algorithm is designed on the DSP processor.The speed measuring system designed in this paper proves that the speed measurement accuracy and speed time of the system basically meet the requirements of the actual use.The function of the modified parameters provided by the system improves the robustness of the system operation and has a high practical value.