Research And Implementation Of Traffic Sign Recognition System Based On Target Detection

The problem of traffic safety in the current society is becoming more and more prominent,and it has become an urgent problem to be solved.Among them,the assisted driving system based on computer vision is an important research direction to solve the problem of traffic safety.Based on this,this paper designs a traffic sign recognition system based on target detection,which can identify and remind traffic sign information and distance information during driving to improve driving safety.The paper conducts research on the target detection algorithm,and combines engineering application technology to design and implement a traffic sign recognition system.The main contributions include:1)The YOLOv5 algorithm improvement method adding CBAM attention mechanism and CA attention mechanism is proposed,so that the network model pays more attention to the key content,and the m AP@0.5 value of the model is increased by 0.6%.An improved method of YOLOv5 target detection algorithm with adding a small target detection layer is proposed to solve the problem that the target detection algorithm cannot detect small target traffic signs well,and the m AP@0.5 value of the model training effect has increased by 5.3%.2)An improved method of YOLOv5 target detection algorithm based on slice reasoning is proposed,the data set is sliced 2*2,and the annotation file is recalculated and the model is trained,which solves the problem that the input image size is too large to effectively detect traffic signs.The detection model performs 2*2 image slices on the input image and performs fusion and NMS suppression on the detected results.The m AP@0.5value of the model training effect is increased by 8.6%.3)Combined with engineering thinking,apply the target detection algorithm,design and realize the traffic sign recognition system,and solve the problem that the detection algorithm cannot measure the distance of the detection target through the binocular camera and the binocular ranging algorithm.Through requirements analysis,system design and other processes,combined with Vue.js,Rabbit MQ,Fast API and other engineering technologies to realize the traffic sign recognition system.Through the research on the above points,the traffic sign recognition system is finally realized,which plays an important role in regulating the driving behavior of users and ensuring traffic safety,and has certain social value and significance.