Traffic Sign Detection Method Research Complex Outdoor Environment

With the development of science and social life, human society has slowly takes on a high degree of information structures. Because of the diversity of information, as well as a flood of information data, it’s very limited capacity to process information, so has the information age is rapidly to intelligent direction. Therefore, development trend of the times helped the emergence of intelligent transport systems. At the same time, traffic sign recognition system as important components of the intelligent transportation systems, its importance and practicability has been widespread accepted, traffic sign detection and recognition of have important representative and scientific merit.This article is intended primarily for urban road traffic sign detection for relevant research, first of all take into account the traffic signs in the outdoor complex environment vulnerable to effects of light and weather factors causing color distortion, made using image enhancement algorithm based on multi-channel correction, respectively, of space and space components are enhanced to improve image color space color distortion problems. Experiments found that this method can effectively restore the color information, enhance the contrast of the image, at the same time by comparing the experimental discovery of HIS space enhancing effects in RGB space. And use components individually and tonal difference of fixed thresholds to color segmentation method for traffic signs, and on the comparison of two segmentation done.Second, when shape detection for the traffic sign, traffic signs use different algorithms for different shapes for testing. For a circular traffic sign detection using an improved transform; triangles, rectangular geometry polygon traffic signs to be detected. Last by experimental tests the validity of the algorithm.Last, on model for has explained, from poly class integrated of angle research traffic flag detection does not degeneration recognition of problem, made a based on visual vocabulary this collective of traffic flag detection method, the method in training image set of interest points set Shang used method to generated different of poly class members, to formed visual vocabulary this collective, and will its application to road traffic flag of detection in the, and are has good of detection effect...