Research On Lane Deviation Early Warning System Based On ZYNQ Platform

With the continuous improvement of China's economic development level,people's living standards are gradually improving.For a more comfortable and convenient life,with using the car as a main tool,the car has gradually entered various families.At the same time,the car bring the increasingly bad traffic problems,traffic has become more and more crowded,and the frequency of traffic accidents has also increased dramatically.Traffic safety has long been a social issue that people in China and the world have attached to.Although the number of traffic accidents has decreased a lot in recent years,the proportion of all problems is still relatively large,and these accidents have seriously affected the family's happiness and social stability.In the related research,nearly half of the traffic accidents due to incomplete statistics are caused by the vehicle deviating from the lane line.If the driver's reaction time is increased,the accident will be greatly reduced.Therefore,it is important to improve the level of lane detection technology,improve the lane departure warning assistance system and reduce the safety of the vehicle during driving.However,in the actual vehicle driving,the road environment is complex and variable,and will be affected by surrounding vehicles.These factors will seriously affect the accuracy of lane line detection.Therefore,the lane line detection system is required to have strong environmental adaptability.In order to obtain high accuracy,the lane line detection still has the problem of large calculation and long time.However,in the vehicle driving,the danger often occurs in an instant,and the lane departure warning efficiency is low,the optimal remedial opportunity will be lost,so the lane is improved.It is of great significance for the line to deviate from the early warning detection efficiency and thus reduce the occurrence of traffic accidents for the public life and social stability.In order to improve the real-time and accuracy of the lane departure warning system,This thesis studies both algorithm processing and system implementation and studies the lane departure warning system based on ZYNQ platform,and obtains beneficial effects:(1)By using the improved Hough and least squares fitting method,the paper can not only eliminate the double edges,but also can stably identify the lane lines of various complex roads ahead,and the recognition accuracy can reach 96.44%.(2)The image preprocessing stage uses a ZYNQ-based system.The experimental results show that the processing power of the ZYNQ architecture is 3.1 times that of the software architecture alone.Compared with pure software design,using FPGA hardware acceleration can greatly speed up the detection process and reduce the time.Overhead,This method can achieve the purpose of real-time detection.(3)Based on the ZYNQ series development platform,this paper studies the problems related to lane departure warning by software and hardware collaborative design,and optimizes the real-time performance of the system.This improves the real-time and robustness of lane departure warning and improves lane departure warning.It is of great significance.