Vehicle driving early warning method and system

Abstract

The invention provides a vehicle driving early warning method and system. The method comprises the steps of: collecting data by a high-speed camera installed in front of a vehicle; performing target detection on the acquired single-frame data; when the detection information is updated, refreshing the tracked historical information; except the first frame, performing target tracking algorithm processing in parallel to obtain a tracking result; performing similarity determination on the tracking result and the detection result of the latest frame of data, and determining whether the target is lost or not; if the target is not lost, updating the historical information; calculating a distance change value between an obstacle and a vehicle; calculating the actual relative distance according tothe latest millimeter wave radar information and the subsequent obstacle distance change; tracking a new frame by using historical information; and performing obstacle early warning based on the tracking result. The method is low in cost; obstacle attributes are accurately determined; the attribute, orientation and distance information of the obstacle is given at a high frame rate; and moreover, the line of sight relates to both front vehicles and side vehicles to avoid jam.

Description

technical field [0001] The present invention relates to the field of traffic technology, and more specifically, to a vehicle driving warning method and system. Background technique [0002] With the popularization of autonomous driving technology, how to ensure the safety of vehicles in fast-speed scenarios has become a difficult point at present. At present, the alarm device mainly uses the relative distance and the speed of the vehicle as the judgment basis. Commonly used in the prior art is millimeter-wave radar and lidar as test tools, or binocular vision. The ranging frequency of the millimeter-wave radar is generally 50ms. Assuming that the vehicle speed is 120km / h in a high-speed scene, the driving distance in 50ms is 1.667m. Moreover, the millimeter-wave radar can only identify obstacles, but cannot judge the risk factor of obstacles. It is difficult to make a quick judgment when there is a sudden vehicle jam. The principle of binocular vision will reduce the fiel...

AI Technical Summary

Problems solved by technology

Based on this, there are at least the following defects in the prior art: the ranging frequency of low-cost devices currently on the market is not high; the attributes of obstacles ahead cannot be accurately and efficiently judged, and correct warning messages are given; The orientation and distance information is uncertain; the camera field of view is limited and the field of view is small

Images