Research On Lane Detection And Lane Departure Warning Method Based On Machine Vision

With the rapid development of artificial intelligence,image processing and other technologies,the environment perception technology based on machine v ision has been widely used;In recent years,lane departure warning by machin e vision has become the focus of intelligent and active safety research;Althou gh there are some breakthroughs in the research of lane departure warning,the re are still many problems.In this paper,the current lane line detection and d eviation warning method is not strong adaptability,detection accuracy is not hi gh,which can’t meet the needs of practical application in a variety of road en vironment to carry out relevant research.This paper proposes a lane line detec tion method that can adapt to various road scenarios,and establishes a corresp onding deviation warning method to enhance the reliability of the whole syste m.This paper mainly does the following research work:Firstly,in view of the shortcomings of the current lane line detection algo rithm’s poor environmental adaptability,this paper focuses on the research of l ane line detection methods in complex road conditions,rain and fog weather,and shadow road conditions.In the preprocessing stage of road image,redunda nt information and noise in the image are removed,lane line edge information is enhanced,and the adaptability and accuracy of the overall detection metho d to environmental changes are improved.Secondly,this paper studies image features including horizontal line,vanis hing point and physical coordinate position,etc.,and then fuses these feature i nformation and combines them with different color models of images and edge detection methods.This paper proposes a lane line detection method with goo d adaptability and robustness,which has better detection accuracy and can effe ctively avoid obstacles and shadow interference in the road ahead.Then,the corresponding relationship between the image imaging principle and different coordinate systems in the imaging process is studied,and the lan e slope detection model is established by combining the prior knowledge of ro ad conditions with the lane line information already detected.The slope detection model is used to detect the actual road,and the test results are analyzed.Finally,the deviation warning method of vehicles on the road is studied.Combining the results of lane line detection with the research status at home a nd abroad,a deviation warning strategy with strong scene adaptability and mor e accurate warning was proposed,and relevant models were established for tes ting and analysis.