Design of vehicle and pedestrian detection and counting system based on YOLOv5
In order to better provide important data support for urban transportation planning,pedestrian flow control and safety management,this thesis designs a vehicle pedestrian detection and counting system based on deep learning.The vehicle pe-destrian dataset is first collected and labeled,and the YOLOv5 model is selected for model training and evaluation of the dataset.Then the trained model is deployed on Jetson Nano 4G core board,which can realize vehicle and pedestrian detection for single pic-ture,video and camera real-time stream,and the detection and counting results are shown through the display.Finally,a user inter-face based on PyQt5 is designed to facilitate user operation and visualization of test results.The test results show that the built sys-tem not only realizes real-time monitoring and accurate counting of pedestrian flow and vehicle flow,but also can support multiple detection methods.
deep learningvehicle-pedestrian detectionYOLOv5 modelmodel traininguser interface
万方数据
