Recognition of triangular traffic signs using the Number of Peaks algorithm

Abstract

Automatic detection and recognition of traffic signs is an important tool in intelligent vehicles. It allows more autonomous vehicles and it can alert the driver to possible hazards and changes in the road. In this paper we focus on the recognition of a wide set of triangular traffic signs using a novel algorithm, the Number of Peaks. Once a traffic sign is detected, three horizontal lines (T, H, B) and three vertical lines (R, V, L) across the image are used to recognize the sign. The number of crossings from a black pixel to a white pixel (peak) on each line is calculated. A simple and fast decision-tree-like search algorithm uses the number of peaks to differentiate between the triangular road signs. A 100% correct detection rate is achievable even in a fairly noisy environment.