Computer vision system for robotic complex based on Arduino and Raspberry Pi

This paper is devoted to traffic sign recognition problem in real time. The recognition process consists of three steps. The first step is a search of image parts which probably contain a traffic sign. The second step is about parts extraction and simple classification by shape. The last step consists of classification of extracted parts with previously learned multilayered neural net. The results of experimental setup based on real data show feasibility of the proposed methods.