Research Of Traffic Signs Detection And Recognition

Autonomous vehicles, also known as autonomous ground mobile platforms,are hot and difficult in the field of intelligent transportation research. Real-timeaccurate perception of the driving environment is the key for the autonomousvehiclesto drive autonomously. This topic targeted on urban road trafficis mainly tostudy the traffic signsdetection and recognition of autonomous vehicles in actualroad. Using saturation threshold segmentation and feature extraction andclassificationwe establish a stable and reliable traffic signs area detection andrecognition system. The main research contents are summarized as follows:First introduce the standard of traffic signs, and lead to the research of thetrafficsigns recognition technology. Through the analysis of the traffic signs imagewe choose the appropriate color space model for preprocessing. According to thecolor characteristics of traffic signs, the feature of the standard deviation ofthesaturation componentgray histogramhas been studied.using F statisticaltest todetect the mutation point of the standard deviation curve, the saturation thresholdsegmentation based traffic signs region extraction algorithm is proposed.On this basis, we went intofeature extraction, and proposed the featureextraction and classificationalgorithm for recognition of traffic signs.Mainly usingHu invariant moments features extraction and SIFT features generating anddimension reduction through PCA algorithm, we design the SVM classifierandachieve a rapid, stable and reliable traffic signs recognition system,which isinvariant for image deformation and effect of shade.Finally, through the test experiments for traffic signsof different styles,different locationsandunder different conditions, we completedthe result analysisand verificationforour algorithm.