A drone-based helmet-wearing detection method, a drone and a medium

Abstract

The application relates to a helmet wearing detection method based on a UAV, a UAV and a medium, wherein the method comprises the following steps: obtaining a helmet wearing detection model based on an improved YOLOv8 network; performing motorcycle detection and helmet detection on a real-time collected road video stream through deploying the trained helmet wearing detection model on the UAV to obtain a motorcycle detection frame and a helmet detection frame; and then determining whether a motorcycle rider in the road video stream wears a helmet in real time through the UAV, and directly reminding the rider to wear a helmet if it is found that the rider does not wear a helmet. Through the application, the improved lightweight YOLOv8 network model is deployed on the UAV with limited computing resources, the detection result is obtained at the UAV end, a warning is given to the rider, the helmet wearing situation of the electric vehicle rider is timely and effectively improved, and the problem of how to improve the helmet wearing situation of road traffic participants by using the UAV is solved.

Images