Research On Multi-target Tracking Method Of Traffic Radar

Intelligent management and control of traffic road is an important measure to reduce traffic congestion.With its unique advantages,traffic radar becomes one of the important sensors in the intelligent traffic control system.It can count the distance,speed,angle,traffic flow and other information of vehicles,realize the detection of vehicle overspeed,congestion assessment and provide effective data for the allocation of traffic lights.Good traffic radar multi-target tracking is the key to ensure the data reliable and effective.Due to the particularity of the traffic radar working system and application scenarios,its multi-target tracking method has many new characteristics and requirements compared with the traditional radar multi-target tracking technology,so it is necessary to study the traffic radar multi-target tracking algorithm.In view of this,this paper designs,analyzes and verifies the whole multi-target tracking algorithm system in the actual scene,and the main work includes the following aspects.1.The principles of range,speed and angle measurement of traffic radar are discussed,and the performance and advantages of traffic radar are analyzed in conjunction with its working system.According to the radar target tracking process,some problems and characteristics that are easy to appear in the traffic radar tracking process are summarized.And conducted a detailed analysis and summary,including its causes,types of problems and solutions.2.This paper studies the multi-target tracking algorithm of traffic radar,including four parts:preprocessing of original track,track initiation,track data association and track filtering.For the first two parts of target tracking:in the pre-processing stage of the original track,combined with the prior characteristics of the road,the identified traffic road edge line is used to screen the target track within the range to be tracked,so as to achieve the effective filtering of the clutter interference on both sides of the traffic road,and reduce the calculation amount of the later track aggregation.Aiming at the problems of fast scanning frequency,large amount of data in a short period of time and no obvious trend of target movement in the early stage of monitoring,an improved track initiation algorithm is proposed for the traffic radar to ensure the real-time track initiation.3.In the process of multi-target tracking of traffic radar,the number of targets is uncertain,and the distance between distant points tends to diverge,so an improved track association algorithm is proposed.When the target track data of a new frame is associated with the track,the track with higher score is preferred to be associated.In the whole process of association,the integrated wave gate with different sizes is set according to the radial distance between the target and the radar to solve the problem of the target track not concentrated in the distance.Since road targets rarely jump significantly between lanes during travel,even if there is a lane change,it is relatively gentle.Therefore,adaptive filtering is used to estimate the successfully associated points,which can accurately reflect the target in time.In recent maneuvers,high-precision and stable tracking of multiple targets has been achieved.4.The realization process of the traffic radar multi-target tracking algorithm is introduced,including the overall implementation scheme and implementation platform of the algorithm.Finally,each process and the whole algorithm are verified and analyzed by the measured data.According to the processing results,the multi-target tracking algorithm as a whole can satisfy the stable tracking of traffic road targets,and can provide accurate and reliable traffic information for the future development of intelligent transportation.