Research On The Traffic Flow Detection Method Based On Radar And Video

Traffic flow and intersection queue length are important parameters in ITS.Millimeter-wave radar plays an important role in security monitoring and traffic monitoring with its advantages of all-weather operation,high-precision velocity and range measurement,and strong environmental adaptability.In the hot era of image processing and machine learning,the importance of video in traffic applications is self-evident.Millimeter-wave radar is not as informative as the information obtained by the camera,but video lacks depth information and will be affected by environment such as light change.This paper focuses on the application of millimeter-wave radar and video in the detection of traffic flow and intersection queue length,and conducts the following research work:(1)Millimeter-wave radar multi-target detection and tracking:the principle of millimeter-wave radar multi-target detection is described,the linear Kalman filter(LKF)and extended Kalman filter(EKF)algorithms are derived,and the simulation is carried out.Through the simulation data and the measured data,it is verified that the observation noise and the process noise at the same time are carried out the adaptive Kalman filter,which will lead to the problem of filtering divergence.Adaptive linear Kalman filter(ALKF)is designed and applied to DSP system.The method to solve the detection velocity ambiguity is improved by using filtering velocity.The proposed method of acquiring intersection queue length based on radar is tested in real-time traffic scene,which can meet the practical application requirements of traffic radar.(2)Traffic flow detection based on video:a statistical method of traffic flow detection based on background difference and a method of intersection queue detection combining edge detection and morphological filtering are designed.The feasibility and accuracy of video traffic flow detection method are verified by the measured data in different scenarios,while the intersection queue detected by video can solve the problem of abnormal jump in the length of the radar detection queue.(3)The specific methods of space calibration and time synchronization of radar and video are introduced.The real-time calibration system of radar and video is designed,which improves the deficiencies in space calibration.The methods of combining radar and video to detect the intersection queue length and the traffic flow are proposed.The measured data verified that the video detection could solve the problem of abnormal jump of the intersection queue length obtained by the radar,and the accuracy of the combined radar and video traffic flow detection was higher than that of the single sensor detection.(4)Based on Visual Studio 2017,WinForms application software is designed under the Windows platform.The software Framework is Microsoft.net Framework 4.5.2,which can be used for radar and video calibration,intersection queue length and traffic flow detection.