The Research Of Lane Recognition And The Methor Of Lane Departure Judgement In ADAS

As people pay increasing attection to driving safety,all kinds of ADAS(Advanced Driver Assistance System)products are appearing in market in an effort to avoid traffic accidents.The lane departure warning system is an important ADAS device as well as a important visual part in intelligent driving which can remind the driver of danger when driver deviates from the lane.Therefore,researchers all over the world has studying it vigorously.This paper proposes a lane departure warning system based on the lane line shape and brightness characteristics.It uses DSP chip TMS320DM6437 as the image processing chip,CCS3.3 as the program development environment,and a MCU development board containing a buzzer as an alarm.The equipment uses the inherent shape and brightness characteristics of lane lines as the basis for road image segmentation.After the lane line detection and fitting algorithm is carried out,the lane departure warning judgement method is finally determined,and a complete lane departure warning system is formed.The general process is that,firstly,choosing the the lower part of the image as the effective area,then preprocess it by graying and median filter.Secondly,the road image is segmented to a binarized image based on the shape and brightness characteristics of the lane line.Thirdly,the center point of all the lane lines in the road image will be extracted based on the continuity feature of the lane line.fourthly,by using the least square method,the line center point is fitted with a straight line model,Finally,the yaw angle and the lateral position of the lane line in the road image are taken as two determination conditions for lane departure and the lane departure warning decision is formed.In order to verify the final results of the system,a real-vehicle test was conducted on the entire system in which the camera installation parameters and the relevant lane departure determination thresholds were determined.The tests show that the system work well in real-time,accuracy and adaptation.