Research Of Detection Algorithm Of Velocity Speeding Behavior Based On UAV Platform

The existing video-based vehicle violation detection technology is based on the pavement fixed camera platform.This paper mainly studies and carries on the research on the algorithm of road moving vehicle detection,tracking and violation judgment,based on the UAV platform.Among them,the most obvious difference between the research method and the traditional method is the difference of video capture carrier.The UAV is a kind of new carrier with some unique advantages,including more flexible operation control,larger monitoring coverage area,lower requirement of the road construction,and the equipment can be unified maintained.At the same time,in terms of the existing traffic management situation in China,it is imperative to expand the coverage of traffic monitoring.Therefore,it can fill the gaps in urban traffic monitoring effectively through the auxiliary management of the traffic vehicle violation detection based on the UAV platform,without the need of large-scale investment,saving costs.At first,Analyzing the situation that the road background in the video capturing by UAV camera is fast moving and dynamic changing,the traditional moving target detection method is ineffective.This paper proposes a background difference method by drift correction based on Mean Shift algorithm,and proves that it is effective on the detection of road vehicles based on UAV platform.it mainly includes: analyzing the objects involved in this study and evaluating various moving target detection methods,after weighing the advantages and disadvantages,takes the background difference method as the basic method of moving vehicle target detection,By optimizing background construction method,proposes the background difference method by drift correction based on Mean Shift algorithm,to realize the target detection of moving vehicles based on UAV platform.Secondly,in the research and realization of the target tracking of moving vehicles,firstly,through a series of image processing methods,including image binarization,filtering noise reduction,morphological filling,shadow removal,gets the clear and accurate foreground moving vehicle target,completing the preprocessing work of tracking.Then,based on the principle of Kalman filter prediction,realizes modeling and Simulation of vehicle tracking by Kalman filter in UAV platform.The vehicle tracking is realized by the algorithm design,and the effect is good and stable.At last,the paper analyzes the method of the vehicle speeding detection,according to the actual characteristics of the video capturing by UAV camera,puts forward the method of vehicle speeding detection of UAV platform based on vehicle trajectory tracking,and proves that this method is feasible and effective by experimental test.