Traffic signs detection, recognition and tracking for roads mapping

Abstract

Traffic signs detection and recognition are well-known problems that are being solved for multiple decades already. There is significant progress in this research area in the last decade. This problem has plenty of solutions that are even deployed in business from driver assistant systems to autonomous vehicles. Nevertheless, we still focus on this area but from a different perspective utilizing the already achieved advances either in traffic signs detection and recognition or in computer vision in general. We want to empower our research group by focusing on the intelligent road infrastructure of which traffic signs are a part. Our contribution to the research group is in composing a solution for real-time mapping of traffic signs. This solution is going to be deployed on real vehicles collecting data about traffic signs like their location, side of the road, visibility, detection distance, or the direction of a road where the sign matters. Finally, this data is going to be crowdsourced by the vehicle agents and collected on a server creating a robust mapping of traffic signs on real roads.