Research On Automatic Identification Algorithm Of Traffic Signs Based On Shape Matching And Analysis

Since the beginning of twenty-first Century,the world economy and the rapid development of transportation,also with the number of cars surge overwhelmed,traffic congestion,traffic accidents have become a major problem in city development.Intelligent transportation system has attracted more and more attention from the government,scholars and scientists are reputed to be the best way to solve the traffic problems,the traffic sign recognition is the key technology of the intelligent transportation system,has attracted many scientists in the field of investment.In this paper,the author summarizes the previous research experience,and takes the traffic sign image of the natural scene as the research object.Mainly introduces the following aspects of work:First of all,according to the images collected from natural scenes of traffic signs,taking into account the changes in image illumination in different angles,different problem shooting distance,the color model conversion,gray normalization,image size normalization of image preprocessing.Secondly,this paper proposes a fast location algorithm based on linear envelope.The use of color information and statistical line information,to obtain the target candidate region,and then use the independent connected region morphology and seed filling method to obtain the label,then use the traffic sign aspect ratio,internal hole information traffic sign out accurately from the target candidate region location.Finally,the shape context matching algorithm is used to identify traffic signs.The first eigenvalue of the covariance matrix of product contour feature points using to estimate the curvature extremum point of curvature,the shape context description thought of distance and angle of feature point information,and then measured by cosine similarity and eight queens optimization matching strategies to identify target traffic signs,traffic signs and database through this acquisition to verify the performance of the algorithm.