Lane position identification method and device

Abstract

The invention relates to a lane position identification method. A plurality of vehicles in front of the vehicle are tracked in the driving process of the vehicle; taking a driving track of a target tracking vehicle which does not change lanes as a tangential coordinate axis of a Fraenet coordinate system of the vehicle; according to the normal distance between each vehicle in the preset area and the tangential coordinate axis, the position of each vehicle in the preset area is clustered, the vehicle group corresponding to each lane can be obtained, and the center line of each lane can be calculated according to the position of the vehicle in each vehicle group. In the running process of the vehicle, the running track of the vehicle running in front of the vehicle is tracked and analyzed; therefore, the distribution condition of the lane center lines on the driving road of the vehicle can be obtained in real time without limitation of network conditions, the lane position can be accurately recognized in real time without depending on a camera to recognize the lane lines and a high-precision map, and the accuracy and reliability of lane position recognition are improved.

Description

technical field [0001] The present invention relates to the technical field of intelligent driving, and more specifically, to a lane position recognition method and device. Background technique [0002] Accurate recognition of lane position is a key link in intelligent driving, and is the basis for realizing functions such as adaptive cruise, auxiliary lane keeping, and lane departure warning. [0003] At present, lane position recognition is mainly realized by using visual perception technology and high-precision map positioning technology. Among them, the visual perception technology is based on the image of road traffic markings (referred to as road lines) captured by the vehicle camera for visual perception, thereby identifying the position of the lane. The road line image that meets the visual perception requirements is captured, so the lane position cannot be accurately identified. On the other hand, high-precision maps are limited by the data collection, verificatio...

AI Technical Summary

Problems solved by technology

Among them, the visual perception technology is based on the image of road traffic markings (referred to as road lines) captured by the vehicle camera for visual perception, thereby identifying the position of the lane. The road line image that meets the visual perception requirements is captured, so the lane position cannot be accurately identified

On the other hand, high-precision maps are limited by the data collection, verification, and update processes. There is a certain period of time between road information changes and high-precision map updates. In the case of road maintenance, temporary traffic control, and traffic accidents, vehicles It is often impossible to obtain high-precision maps that can reflect road information in real time. Therefore, in this case, positioning technology based on high-precision maps cannot accurately identify the lane position

[0004] In summary, the existing methods for identifying lane positions are poor in accuracy and real-time performance, and there is an urgent need for a method that can accurately identify lane positions in real time

Images