The Technique Research On Camera-based Lane Line Recognition And Lane Departure Warning

"Made in China 2025" clear,by 2020,the preliminary independent research system and the production supporting system for intelligent automobile should set up in our country.To speed up the cultivation and development of intelligent and connected automobile is a wise strategic choice.It complies with the change of the global auto-industry's development trend,helps to grab the commanding heights of the auto industry.Lane departure based warning technology is one of the important basic technology of the first stage of the intelligent and connected automobile.To strengthen the research of lane departure based warning technology has very important theoretical significance and practical application value for realizing the intelligentizing of the vehicle driving.An extensive investigation on the current situation of the development of the intelligent and connected automobile has been completed,and so has a deep analysis on the current situation of camera based lane line identification and lane departure warning technology.Then,based on that,this paper chooses the city-road surface that contains complex interference for research.Main contents include the recognition algorithm for lane line and warning algorithm for lane departure.The key technology of the two algorithm has been carefully studied and discussed.The main research work that has been completed and the related research results are as follows:1.This paper proposes a method of the lane line recognition.As the current recognition methods for lane line cannot effectively recognize yellow lane line faults,this paper proposes a new method of the lane line recognition.First,the yellow lane line in the photo is recognized in the light of the pixels' brightness of yellow lane line is low but their yellow and edge feature are obvious.Then,the white lane line in the picture is recognized by using threshold segmentation method according to the characteristic that the white lane line pixel is brighter.Finally,the result of the yellow and white lane line recognition is obtained by merging the two recognition results.2.The camera model has been established and the internal and external parameters of the camera have been calibrated.First,in order to calibrate the internal and external parameters of the camera more easily,a pinhole camera model has been established.Then,based on this model,the internal and external parameters of the camera have been calibrated separately.3.A warning method for lane departure has been proposed.First,an in-depth comparison of the current wide adopted warning methods for lane departure is made.In addition,these methods' applicable occasions and their advantages and disadvantages are analyzed.Then,a new warning method for lane departure is proposed by combining the best of future departure time based TLC method and current position based CCP method.This new method is called the future position based CCP method.In order to prevent unnecessary false alarm from interfering the driver,this paper also designs the alarm-blocked algorithm.