Traffic Sign Recognition Research&Implementation On Mobile Devices

With the development of the economy and the popularity of car, traffic accident problem is more and more serious. As important source of information on the road, traffic signs can provide traffic information and limit the behavior of traffic participants.As part of the ITS, traffic sign recognition (TSR) has extensive research and application prospects. Scholars in the field of TSR have made abundant achievements, but most of the research is based on analysis of traffic signs in static image. Given this background, this paper studied, used and optimized the preferred choice of algorithms to build a TSR system running in iOS.This paper selected and optimized the algorithm based on color and shape to detect traffic signs:using the HSV color space, system realized the segmentation of traffic signs quickly and accurately; using morphological methods, system eliminated noise and got the region of interest; using Hough transform, the shapes of traffic signs were detected. In addition, simplified LeNet-5convolution neural network system was used to recognize traffic sign. At the end of this paper, the algorithms selected were implemented in iOS.This paper used the traffic sign database and images collected in real scene to test the system in a variety of weather and light conditions. It can be seen from the results, in this paper, that the selection of algorithms and application built in iOS has good robustness.