Lane line detection method and device

Abstract

The invention provides a lane line detection method. The method comprises the steps that real scene images are acquired; inverse perspective transformation is performed on the scene images so that aerial-view images are acquired; the gradient of the aerial-view images in a horizontal direction is calculated; gradient width filtering is performed so that candidate line segment discrete points are acquired; clustering of the candidate line segment discrete points is performed by using a clustering algorithm; feature point screening of the discrete points of each cluster is performed so that feature points are acquired; fitting of the feature points is performed by a least square method so that candidate lane lines are acquired; and screening of the candidate lane lines is performed and the lane lines are outputted. The lane lines under the complex scenes can be greatly detected.

Description

technical field [0001] The invention relates to image processing, video monitoring and driving safety, in particular to a lane line detection method and device. Background technique [0002] With the gradual development of the transportation and vehicle industry, the number of vehicles is increasing day by day, and the annual traffic accidents are also increasing sharply. According to statistics, the direct or indirect causes of road traffic accidents include people, vehicles, roads and various environmental factors, and 44% of all fatal traffic accidents are related to lane departure. [0003] The Chinese invention patent application with the publication number CN103383733A discloses a lane video detection method based on semi-machine learning, which establishes a background image, uses median filtering and edge detection for processing, and uses improved Hough transform to detect straight lines. The K-means algorithm implements clustering, using the straight line of the c...

AI Technical Summary

Problems solved by technology

However, the above methods have a low la

Images