Detection method of virtual and real lane lines in traffic monitoring scene

Abstract

The detection method of virtual and real lane lines in traffic monitoring scenes generates lane line images based on the surveillance video of multi-lane roads, and sets the lane ROI; then detects the lane lines in the lane ROI, performs clustering, and classifies the lane lines based on the clustering results. After classification, based on the geometric features of the lane line, divide the left and right edges of the line segment of the lane line, and extract the endpoint information of each virtual lane line for the virtual lane line; The lane line is fitted to obtain the final lane line detection result. The present invention is mainly aimed at multi-lane monitoring scenarios, and can effectively solve the problems existing in traditional lane line detection methods in the multi-lane scene in terms of lane line occlusion, left and right edge division of lane lines, and accurate information extraction of virtual lane lines. It has important and far-reaching significance for applications such as automatic driving and automatic calibration based on accurate lane information.

Description

technical field [0001] The invention belongs to the technical field of computer vision detection, relates to the analysis of multi-lane traffic monitoring video, and provides a method for detecting virtual and real lane lines in a traffic monitoring scene, which detects and extracts complete lane line information in a multi-lane scene. Background technique [0002] Lane line detection is one of the current research hotspots, and it can serve the fields of automatic driving, automatic calibration, and video-based traffic monitoring. However, there are still some challenges in the current lane detection, such as lane occlusion in multi-lane scenarios, left and right edge division of lane lines, and endpoint information extraction of virtual lane lines. The application of lane line information in intelligent traffic monitoring puts forward higher and higher requirements for the accurate detection of lane lines, but traditional algorithms are difficult to meet. At present, most...

AI Technical Summary

Problems solved by technology

However, there are still some challenges in the current lane detection, such as lane occlusion in multi-lane scenarios, left and right edge division of lane lines, and endpoint information extraction of virtual lane lines.

The application of lane line information in intelligent traffic monitoring puts forward higher and higher requirements for the accurate detection of lane lines, but traditional algorithms are difficult to meet

At present, most of the existing lane line detection algorithms are aimed at intelligent driving scenarios, and they are all based on the image of the road conditions in front of the vehicle. There are few detection methods suitable for multi-lane road monitoring scenarios, and the lane line information obtained by intelligent driving detection cannot be completely Meet the needs of traffic monitoring

Images