Research On Traffic Target Tracking Algorithm Based On Radar And Video Fusion

In the Intelligent Transportation System,in order to obtain effective traffic information for analysis and processing,it is necessary to track the vehicle target accurately.Different sensors have advantages and disadvantages.As one of the commonly used sensors in the transportation field,millimeter wave radar can detect the parameters such as the distance,speed and Angle of the vehicle target.However,the millimeter wave radar cannot get the shape and detailed information of the vehicle.Cameras operate on various roads in the city,However,the camera sensor is greatly affected by light,rain and snow,and lacks speed information.Multi-sensor fusion technology,as one of the important components of building a smart transportation system,has gradually developed in recent years.Therefore,this paper proposes a traffic target tracking algorithm based on radar and video fusion,which can solve the problem of multi-target tracking in complex traffic scenes.The main research contents are as follows:(1)Process radar and camera data to obtain target tracking results.Radar data target tracking adopts adaptive filtering,output target ID,distance and speed.Camera data target tracking adopts Kalman filter,output target serial number and centroid.The results are used as input data and comparison data for the fusion algorithm.(2)Due to the different data rates of radar and camera,it is necessary to perform spatiotemporal matching of sensors.For time matching,align the time starting point of the data and the data frame rate.For spatial matching,the conversion between five coordinate systems.By obtaining the internal parameters of the sensor and the spatial information when the device is installed,the data in the radar coordinate system is converted to the pixel coordinate system of the video.The matched data can be displayed in the same dimension,ensuring the accuracy of data fusion(3)For the problem of diversified target movement in complex traffic scenes,the multisensor fusion algorithm of fuzzy theory based on decision level is chosen as the main algorithm architecture of this paper.According to the multi-radar point problem,the video clutter problem and the target segmentation problem,a target decision criterion is proposed to improve the traditional fuzzy set theory data fusion algorithm,and the Z-type membership function is introduced to optimize the decision branch.The improved fusion algorithm can solve the problem of radar more,video clutter and target segmentation,It can improve the tracking accuracy on the basis of the traditional algorithm.(4)The fusion tracking algorithm is verified by collecting a large amount of measured data of traffic roads.The results show that compared to the single-sensor target tracking results,the multi-sensor fusion algorithm can solve the problems of multiple radar points,video clutter and target segmentation.Its tracking rate is better than the single-sensor tracking rate,and it is more stable in complex traffic scenes.It can provide richer and more accurate information for the development of smart transportation system.