Research And Implementation Of Car Lane Departure Warning System Based On Image Processing

With the rapid development of our country's economy and the improvement of people's living standard,car ownership increases rapidly in recent year.Thanks to the increasing number of cars,the traffic environment is becoming worse which not only leads to traffic jams but also more and more traffic accidents.In order to reduce the rate of traffic accident,the intelligent driver assistance systems are invented,and the Lane Departure Warning System(Lane Departure Warning System,LDWS)is the core of those systems.Lane Departure Warning System detects the road images which captured by the camera to obtain the current position information of the vehicle.It reminds the driver through warning signals when the vehicle becomes to deviate from the lane.This article analyzes the domestic and foreign lane departure warning systems at first.The content of this article is mainly divided into three parts which include image preprocessing module,lane detection and tracking module and lane departure warning module.The main work is as follows:1.Image preprocessing.It consists of gray processing,filtering enhancement,image binarization and edge detection.Through analyzing several different kinds of image preprocessing algorithms,the suitable method which is combined with the actual needs of road images is chosen firstly.Then,the linear model is chosen as the lane model.Through analyzing kinds of linear detection method,Hough transform is selected.2.In the process of lane detection,a lane detection method which combines hough transform and gray block matching is proposed.It works as follows: First,images are detected several lines by hough transform.And then,comparing with the different between road gray and lane gray,the actual lane is detected by gray block matching.3.Through analyzing kinds of classical lane departure warning system model,the model which combines lane departure rate and linear number is proposed.Firstly,the lane departure rate is calculated by lane angle.Then,the current state of the vehicle is judged by adding the condition of linear number detection,which aimed to meet the needs of warning precisely.Furthermore,the different conditions of lane departure is analyzed.According to the test,the lane detection method and lane departure warning method proposed by this article has a significant effect.As well as the LDWS is very effective.