Research On Lane Recognition And Departure Warning Algorithm Based On Hyperbolic Model

Nowadays,with the rapid development of the world automobile industry and the increasing of vehicle population.Safety problems have become more and more prominent for the high incidence of traffic accidents,especially the serious accidents.The intelligent driving assistant system arises at the historic moment,which can provide useful warning information to driver before lane departure and improve the driving safety greatly.This article embarked from the vehicle active safety,and carried out active exploration in the field of intelligent vehicles.The road recognition,lane tracking,departure warning algorithm and experiments was studied deeply.To improve the accuracy,reliability and computing efficiency of lane recognition and departure warning algorithm,a novel lane detection and departure warning framework based on the hyperbolic model was proposed.The color space transformation,threshold segmentation and other steps were used to reduce processing time.In lane recognition stage,lane edge points were obtained from preprocessed image by searching feature points and particle filter algorithm to improve the search efficiency.A least square fitting method was used to identify hyperbolic model of lane.The confidence factor of identified lane model was evaluated by a confidence function,which improved the robustness of recognition algorithm.Finally,lane parameters were obtained by camera calibration transformation matrix,based on a hyperbolic lane model established in aforementioned procedure,a spatial and temporal warning model of lane departure was proposed in world coordination system to reduce the false alarm and missing alarm rate of the system.To verify the validity of the proposed road recognition and departure warning algorithm,the lane departure warning model was built on the co-simulation platform using PreScan and Simulink for simulation tests,and the algorithm was realized by means of VC and OpenCV to do road experiments.Simulation and experimental results show that the method possesses good performances in precision of recognition and computation efficiency,and can meet the application requirements.