A Method for Recognizing and Ranging Pedestrians and Non Motorized Vehicles Based on Monocular Vision

In order to avoid accidental spraying of pedestrians and non motorized vehicles by sprinkler trucks during operation, a method for identifying and ranging pedestrians and non motorized vehicles based on monocular vision is proposed. Firstly, the Zhang calibration method is used to calibrate the parameters of the monocular camera, obtaining the camera parameters. Then, SEAM is added to the YOLOv8 (You Only Look Once version 8) algorithm Attention mechanism and improved YOLOv8 algorithm are used to detect pedestrians and non motorized vehicles, ByteTrack algorithm is used to track targets, and finally the target distance is obtained through monocular ranging module. The experimental results show that the improved YOLOv8 algorithm has improved the precision, recall, and mAP50 of the original model by 0.16, 0.17, and 0.05, respectively, and can effectively measure the distance between the front of the vehicle and the target distance.