Research On Vehicle Flow Monitoring And Vehicle Recognition Technology Based On Magnetic Field Detection

The detection and recognition of vehicles in intelligent transportation systems is an important part.The collection and upload of road traffic conditions can be completed efficiently,accurately and at low cost,which can greatly improve the current overall traffic decision-making efficiency and the traffic experience of traffic participants.This article mainly studies a traffic information monitoring system based on TMR single-axis magnetic sensor.Compared with traditional traffic information acquisition methods such as cameras,lidars,and buried coils,it has low cost,less weather influence,easy installation,Feature with little impact on real-time traffic.This paper designs the overall software and hardware of the vehicle detection system’s identification system.The hardware part includes power management module,front-end TMR sensor acquisition module,analog signal processing module,MCU module and communication module,etc.;software includes overall vehicle monitoring multi-task system,communication protocol system,multi-lane master-slave mode setting,etc.In addition,a host computer interface that can perform network communication or serial communication with sensor nodes is also designed,which is convenient for realtime display of the magnetic field intensity waveform and traffic volume of the detection area,the number of different vehicle types,the average speed of traffic volume and other traffic parameters.In terms of vehicle detection algorithms,a variance sequence state machine based on magnetic field strength values is designed to monitor vehicle flow,which overcomes the drawbacks of using only magnetic field strength sequences or variance sequences for vehicle detection,and achieves high-accuracy vehicle flow monitoring effects and data Based on the segmentation algorithm,it realizes the estimation of the vehicle speed,the determination of the vehicle direction and the tracking of the reference value of the magnetic field environment,and provides a high-quality data set for the vehicle type recognition algorithm.In the vehicle recognition algorithm,the feature extraction of the vehicle magnetic field waveform is completed,and the feature dimension reduction is performed based on the PCA algorithm.The centroid separation algorithm based on the K-means clustering algorithm and the SVM classification algorithm based on the radial basis function are designed to achieve high accuracy classification of small cars,medium cars,trucks and buses.The centroid separation algorithm separates buses from other types of vehicles,improves the speed of the classification algorithm,and reduces the burden of the subsequent SVM algorithm model.This paper conducts road field experiments on the designed sensor nodes,corresponding embedded software,vehicle detection algorithms and vehicle identification algorithms.The experimental results verify that the system is stable and can meet the needs of obtaining road traffic parameters.