Study On Lane Mark Recognition In Lane Departure Warning

Statistics show that the majority of traffic accidents are caused by careless negligence of the driver.When the vehicle deviates from the lane marking while driving,the lane departure warning system can immediately warn the driver so that the vehicle can run normally,thereby preventing traffic accidents.The first technical problem to be solved in this system is lane markings recognition.Therefore,this paper conducts a comparatively in-depth study on lane markings recognition technology.Lane markings are the most basic signs identified by the car during driving.The research on lane marking detection methods in lane departure warning systems is one of the research directions of automobile-assisted driving systems.This paper proposes a specific detection method for structured roads.In visual navigation,there are very high requirements for the accuracy of road monitoring and road positioning.In order to realize the above requirements,this paper studies the recognition method of lane markings in lane departure warning technology,and introduces the image processing flow and method of lane marking recognition in detail.It uses Matlab to perform various processing algorithms for each module.The simulation comparison and analysis improved the maximum between-class variance binarization method and Canny edge detection algorithm,improved the recognition effect,and carried out an example simulation analysis according to the proposed algorithm flow.This paper analyzes the image of lane markings collected by machine vision.For each extracted frame of image,the image preprocessing algorithm of weighted average graying and median filtering is first used to filter out the interference information in the lane line image,reduce the data that the image needs to process,simplify the subsequent analysis and processing,and then use the histogram equalization method to enhance the gray-scale contrast of the lane line image,and extract the ROI from the region of interest of the image,adopting the improved maximum variance between classes and the improved algorithms such as Canny edge detection improve the recognition effect of lane markings.Finally,the Hough transform is used to fit the lane markings according to the characteristics of the lane markings,and the recognition of lane markings in the images is completed.The verification results show that the algorithm flow used in this paper can handle the lane markings image well,detect the edge of the lane markings accurately,and then complete the identification of the lane markings.It has good stability and real-time performance.