Searching and recognition of road signpost using ring detection network

Abstract

We propose a recognizing system of speed limit signs in a driver's front view by image processing. At first, in the preprocessing, the input image from the camera is processed by the contrast and extraction of the colors of road signs using color chart. Secondly, the position and the size of the object sign are detected by the ring detection network. The output is calculated by taking the product of those picture element values and matching-pattern values. Matching-pattern is shaped circular ring of different radii radius. Each output value corresponds to the size of the signpost. From this result, the existence and size of signpost is detected. The figure pattern is cut out and reconstructed using interpolation method, assumed to consist of 8 length and 5 dots. The recognition objects with three kinds of speed limit sign 30, 40, and 50 are considered with each kind of speed limit sign, 100 times of experiment is carried out with a total of 300 experiments using neural network. The success rate is 98.3%. This method can detects only the ring. The application can be expected of many fields of the image processing. Moreover, these processes are carried out using networks by many connecting weights. This system has the possibility to implement hardware system using parallel processing technology, and is able to realize real time processing.