Research On The Street Lamp Positioning System Based On Lamp Pole Recognition And Trajectory Tracking

Street lamp cleaning is the infrastructure of road traffic construction.Cleaning street lights regularly can ensure people’s travel safety.At present,the work of street lamp cleaning mainly relies on manually,which poses a safety problem for staff members.This paper starts from an automatic street lamp cleaning device,and develops the street lamp positioning system based on target lamp pole recognition and tracking.The main research is as follows:(1)The overall design of the scheme of street lamp positioning.The installation parameters of the street lamp are determined by analyzing the working environment of the cleaning device.In this paper,a street light positioning scheme is proposed,according to the Light Pole and guardrail.The angle between the cleaning device and the guardrail is obtained by the ultrasonic sensor,and the position of the light pole is obtained by the laser radar,and the height of the cleaning platform is obtained by the single point Laser rangefinder.Combining with the physical parameters,the central position of the Lampshade is determined.(2)The recognition of target lamp-post based on the physical characteristics.According to the radar scanning range and data characteristics,the data processing amount is reduced by dividing the region of interest.The hierarchical clustering algorithm is used to cluster the point cloud data,and the clustering threshold is determined by experiment.Gaussian filtering is used to complete data smoothness processing,and the results show that the filtered data is smoother and more consistent with the arc trajectory.The least square method is adopted to realize the arc curve fitting,and the target lamp pole is identified through the characteristic value of the lamp pole.In order to enhance the reliability of light pole identification,according to the principle of least squares method,the average square sum error is introduced as the basis for light pole identification.(3)The research on the target of street light pole trajectory tracking.Aiming at the movement characteristics of the cleaning device that are slow and the rate of speed change is low,the Kalman filtering is used to track the target light pole trajectory.The movement scene of the cleaning device under different working conditions is designed,and the rationality of the algorithm is verified by simulation experiment.The results show that the Kalman filter and the motion model of the street lamp cleaning device meet the requirements.(4)The experimental verification of street lamp positioning system.The impact of data smoothness on target light pole recognition is analyzed,and the results show that the fitting radius of the processed data is closer to the real radius of the lamp pole and the square sum error is smaller.The feasibility of two lidars to identify the same light pole is tested.The data of Lamp Pole are collected from different distances,and then the date is transformed to the same radar coordinate system and fitted.The results show that the identification of the same lamp-post by two lidar is more robust.The feasibility of the street light positioning system is verified.The experiment platform of street lamp positioning system is built and the experiment of street lamp positioning is verified.The vehicle is parked under the street light through the street light positioning system,and it proves the feasibility of the street lamp positioning system.