Research On Detection And Recognition Of Road Traffic Signs

Road traffic signs detection and recognition system is one of the core technology of the intelligent transportation system, and it is the main component of the automatic driving assistance system. Research on detection and recognition of road traffic signs can improve vehicle traffic safety, it also can promote intelligent vehicle development process. So the research have important significance and application value.Research on detection and recognition of road traffic signs is based on traffic signs in complex background. Shooting traffic signs environment have different conditions,such as point of view,light,weather,location and so on. There are some problems such as traffic signs color distortion, shape change, noise and son on. This article use image graying, image restoration and image edge detection method to reduce image interference, it also can highlight the region of interest. Technology research to be able to hold detection and recognition of road traffic signs accurately. And system can provide tips,guidelines,warning and other traffic sign information to autonomous vehicles. Thus, the paper select Canny operator to extract edge information,and it select RHT algorithm to detect road traffic signs.Using HOG feature extraction methods to obtain valuable characterization. The recognition of road traffic signs combined with linear kernel based SVM.Results of this research show that the RHT algorithm compared with SHT and other algorithms, the accuracy of the test results have an advantage. Compared with other feature extraction algorithms, the HOG feature extraction has a high time consumption, but it has a higher classification accuracy. The recognition accuracy of SVM based on linear kernel function is better in dealing with the problem of traffic sign recognition under the complex background of interference factors.