Research on urban road safety early warning system

Abstract

Based on AI, big data, the Internet of things, edge computing and other technologies, providing perfect, rich, and professional software and hardware solutions can further promote the development of the transportation industry and improve the informatization and automation level of intelligent transportation. LDWS is a system that assists drivers in reducing traffic accidents caused by lane departure by means of alarm. When the vehicle exceeds a certain speed, it can send out an alarm before unconsciously (the driver does not turn on the turn signal) departing from the original lane, providing more reaction time for the driver, and greatly reducing the collision accidents caused by lane departure. It detects the lane line in real-time and estimates the departure time according to the steering wheel direction, vehicle speed and the angle between the vehicle and the lane. If the departure time is less than a threshold, it will send an alarm to remind the driver. The lane departure warning method based on computer vision technology uses a monocular camera to calibrate the lane line and obtain the holography matrix of the world object plane and image plane; Then detect, track, and calculate the slope of the lane line in the obtained lane image, and judge the lateral displacement of the vehicle and the change of the lane line slope. Through optical fiber 5g, AI vision chip and artificial intelligence algorithm, artificial intelligence network calculation, logic operation and embedded system operation are realized, and applied to vehicle deviation early warning system, intelligent speed measurement early warning system, curve passing early warning system, traffic accident early warning system, intersection early warning system, intelligent traffic video perception system and traffic accident early warning system. The early warning system supports the import of a control database, can generate data reports, and provide edge operation support for the solution through traffic flow thermal map distribution imaging and traffic flow distributed counting. It can accurately capture and identify the machine, non-machine, human and other elements, and master the road traffic information in real time. Use the data to analyze the road conditions, provide accurate video image data for the traffic management department, and guide the traffic system to divert and adjust the traffic information in time in combination with the flow data. At the same time, real-time detection of road status, number of pedestrians, moving track and other information, prediction of risk coefficient, intelligent reminder, active broadcast, and early warning, and update the early warning content at any time.