Reseach On Lane Departure Warning System Based On Machine Vision

In recent years,along with the rapid development of the society and economy,the number of automobiles,such as motor vehicle,and the total length of roadway is increasing rapidly,which also result in high frequency of road traffic accidents.Therefore,how to improve the safety factor while driving and reduce vehicle traffic accidents have become the hot topics in the research field of Automobile Safety.Lane Departure Warning System(LDWS)is of vital significant for reducing accident caused by Vehicles' improper lane changing.Besides,LDWS is an integral part of driving assist technology,automobile safety technology and self-driving technology.At present,due to those factors such as high cost,big volume and poor real-time capability,the research on LDWS is still in its initial development phrase.In this paper,in view of the problems existing currently,hardware design and algorithm design have been proposed which based on present research as well as necessities of LDWS.The hardware chapter mainly consists of signal acquisition module,data processing module,memory module,and alarming & displaying module.Futher explanation of those modules' functional properties and schematic diagram designing have been given.Forthermore,detailed introduction of driver program design for some modules has been given also.In the algorithm design part,firstly,the simulation for filtering algorithm and edge detection algorithm have been verified,and the transplants of algorithm to the hardware platform were finished,which are based on above simulated results.Then,heuristic searching algorithm was applied to design a kind of optimized boundary point searching method.In addition,Hough transformation was applied to finish lane boundary point curve fitting.In order to make Hough transformation more efficient and realtime,meantime,takeing those defects of Hough transformation such as time-consuming and inefficiency into account,a spatial parallel algorithm was proposed.Lastly,a Lane departure warning model for horizontal positions and deviation angle was accomplished,which were based on lane line extracted in advance.In this paper,a complete Lane Departure Warning System which include hardware design,majorizationand transplant of algorithm have been designed.Finally,the functional test results of the platform showed that the expected effects have been obtained.