Design Of Lane Line Recognition And Departure Warning Based On Structured Road

Intelligent transportation has been a focus of attention in recent years.Unmanned and assisted driving in the field of intelligent transportation has facilitated people's travel,while reducing traffic accidents caused by driver fatigue and road congestion.Lane line detection technology and vehicle departure warning technology are two crucial technologies.These two technologies can detect and identify lane lines in real time,and determine the specific location of the lane where the vehicle itself is located.Finally,it is fed back to the terminal to remind the vehicle to drive within a safe range.Therefore,it is of great significance to study the lane line and vehicle departure early warning system accurately and in real time.In order to improve the accuracy and real-time performance of lane line recognition and vehicle departure warning,and to timely feedback the driving conditions of the vehicle to the driver.This thesis first systematically researches various methods of lane line recognition based on structured roads.Secondly,the characteristic attributes of lane lines are analyzed.Finally,the vehicle departure warning information is sent to the terminal,which effectively reduces and avoids traffic accidents.The main research contents of the paper are as follows:(1)A system frame for lane line detection is set up.The recognition and detection of lane lines are studied.Real-time detection and marking of lane line position.Due to information redundancy of video data,noise interference,etc.First,the data is preprocessed by extracting regions of interest,graying,filtering,and binarizing.Then research and compare multiple edge detection algorithms.Canny edge detection operator was chosen to perform edge detection on the processed binary image.Next,Hough straight line detection is performed on the detected edge information to obtain all straight line information data in the figure.The effective lane line information is obtained by filtering all straight line information according to the characteristic attributes such as lane line length,slope,position,and color.(2)By studying the characteristic attributes of lane lines,a system model of vehicle departure warning is derived.Firstly,four classic basic deviation early warning models of FOD,CCP,TCL and KBIRS are analyzed.We make improvements on this basis and come up with a vehicle departure warning model based on slope.This model is divided into two levels of judgment conditions.The first layer uses the slope of the lane line as a judgment condition.The second layer uses the relative distance of the vehicle center to the lane line as the judgment condition.Through the above two layers of judgments on the detected lane lines in order,the current specific deviation of the vehicle can be finally obtained.(3)The lane line detection framework and the vehicle departure warning model are combined to complete the system design and applied to multi-scenario video data.Good weather videos,night videos,and rainy videos have their own characteristics.The light around the night video is too dark,and the overall clarity of the rainy video is low.The system in this thesis can have relatively good detection results.