Prohibition Sign And Warning Sign Detection Technology

In this paper, traffic signs detection methods in the natural scenes are studied, and then a real-time traffic signs detection algorithm is summarized. The main contents are as following:Color segmentation for prohibition and warning signs is studied. Besides the HSV color space segmentation and RGB color space segmentation, the simple vector filter SVF color segmentation is studied. The RGB space color segmentation is instability and vulnerable to climate and scene conditions. That the space transformation from RGB to HSV is nonlinear leads to a large amount of calculation and time-consuming, which is not conducive to real-time sign detection. The simple vector filter has simple principle that is use achromatic factor to extract traffic sign colors. There is some improvement in the simple vector filter color segmentation. After extracting chromatic color through the simple vector filter, the three components corresponding of color are used to extract color further.There are many shape detection methods on prohibition signs. The HOUGH transform, round-degree method, and the corner detection method are the main circle detection methods. The HOUGH transform circle detection method involves complex computation, and needs to set some restrictions in advance, which have certain limitations for real-time traffic signs detection. The round-degree method is simple and easy to implement, which only use the round-degree to measure traffic shape. In improved round-degree method, the regional geometry width, height and geometric aspect ratio are introduced. Use these regional geometric parameters to segment region coarsely, which will help improve the detection efficiency.Extracting feature parameters of triangle is applied to detect triangle traffic sign. And compared with corner detection, applying the triangle-degree to detect triangle sign saves computing time and is conducive to real-time detection.For real-time traffic signs detection system, applying tracking and detection method can improve the detection efficiency. It is mainly studied on KALMAN filter tracking. The tracking result is adopted to detection traffic sign. The KALMAN filter is improved to realize multi-target tracking and detection.Experimental results show that the proposed color segmentation and shape detection algorithm for the prohibition signs and warning signs have very good applicability. The method saves time for traffic sign detection and improves the accuracy of traffic sign detection.