Research On Traffic Sign Recognition Based On Vehicle Vision

With the development of economy and society, taking the automobiles as thedominant traffic tools grow with each passing day, Road as a carrier of thepassage of motor vehicles,containing traffic information in which has a importantinstruction function no matter to the human driving or unmanned vehicle。This paperstudies on the part of road traffic information and put forward a new method to solvesome key technical problems in traffic scenes。Because of the natural scene under theroad traffic is extremely complex, which makes the quick find signs the area where itis particularly important, the establishment of the region of interest is one of theeffective ways to solve the problem。By using the inverse perspective transformationcould not only establish the desired region of interest, but also the lane line in theimage information to be correct, it provides convenience for subsequent markerdetection and lane detection。In order to research the road traffic identification Better, this paper proposes asegmentation method based on global average gray and local average grayapproximation of the optimal threshold, and this method was compared with the Otsumethod and the Iterative Threshold method by using standard image. Theexperimental results show that our method proposed in this paper is robust。In order todetect zebra crossing in central of road, based on the shape features of crosswalk,detection method is proposed in this paper based on the horizontal crosswalkprojection, and using the experimental results to illustrate the effectiveness of themethod。...