Image-based Lane Recognition And Research On Automobile Lane Departure Warning

In recent years,the number of car ownership in China has increased year by year,and the traffic safety of automobiles has received more and more attention from the society.In order to reduce the number of traffic accidents and strengthen the active safety of the car,the intelligent assisted driving system is gradually equipped.Lane departure warning technology is one of the core technologies of advanced driver assistance system.In high-speed scenes,due to driver fatigue,drowsiness and distraction,if the vehicle deviates from the lane,the system will give the driver a warning signal in time to avoid collisions.The occurrence of traffic accidents ensures the safety of the driver's life.This paper mainly studies image-based lane recognition and car lane departure warning technology.The core modules of this technology are lane recognition and tracking algorithm and lane departure warning algorithm.Firstly,the development status of the lane departure warning technology at home and abroad,as well as the research status of the core algorithms that make up the technology are introduced,and the existing algorithm deficiencies are pointed out.Then the detailed design process of the three core algorithms involved in this paper is expounded.The main contents of the algorithm are as follows:1.Preprocessing of road images.The road image sequence taken by the front camera is pre-processed frame by frame.The image preprocessing process mainly includes setting the region of interest,graying,median denoising,top hat enhancement processing,threshold segmentation,contour restoration,contour center feature extraction,etc.By image preprocessing,noise interference can be removed,target lane characteristics are enhanced,and preparation for subsequent identification detection is prepared.2.Identification and tracking of lane lines.The expressway has a large turning radius,and the image near field of view can be simplified to a straight line,and the curve model has a large amount of calculation and poor real-time performance.Therefore,a simple straight line model is selected to describe the lane line.In order to improve the accuracy of the initial frame,the RANSAC algorithm is used to fit the feature points of the road contour center of the initial frame,and the Kalman filter is used to track the lane line parameters,and the left and right sides are deviated by acertain distance to narrow the region of interest of the tracking frame.Since the tracking process reduces the noise interference of the image,the least squares method is used to fit the contour center point of the region of interest.3.Image-based lane departure warning.The traditional lane departure warning algorithm requires a complex camera calibration process to determine its parameters.This paper proposes to determine whether the vehicle is in a deviated state based on the number of lane lines and the angle between the images.First,the number of lane lines is detected.If the double lane line is detected,whether the vehicle is deviated is determined whether the difference between the angles of the lane lines is greater than a set angle threshold;if a single lane line is detected,a single lane is utilized The empirical value of the line angle is determined.And the effectiveness of the early warning algorithm based on the number of lane lines and the angle is verified by the real shot video of the vehicle.The software for lane departure warning system is realized by Matlab 2014 a programming,using Matlab's file interface to read video image sequences,processing and marking frame by frame.Simulation experiments show that the proposed lane detection and tracking algorithm has good accuracy and real-time,and the effectiveness of the early warning algorithm.