| In this thesis, a new method for detecting and recognizing road signs from real world scenes is presented. The main outline of this method is composed of three parts: detecting road signs, i.e. road sign segmentation, shape recognition of segmented areas, and identifying the meaning of the detected signs in a hierarchical method. The segmentation of the road signs using a simple edge detection algorithm may cause many false line extractions on the signs. Instead, a texture algorithm, an RG, and a BY color opponent image are proven to be a very effective way to segment the signs due to the peculiar characteristics of a sign's surface and color against its background.; After the segmentation, the shape of the extracted sign is identified by a few grid lines and a bounding box surrounding the sign. The shape of the sign gives a hypothesis for searching the possible matches in a modelbase in a hierarchical way. In addition, the color information of the signs taken from the segmentation phase can help subdivide the hierarchical method. The modelbase forms compressed eigenspaces of signs for each corresponding shape; i.e. triangle, diamond, rectangle, pentagon, octagon, and conic. The eigenspaces of the modelbases are used for the identification phase along with the eigenspaces of the detected signs. |
