Triangular traffic sign recognition
Abel Sánchez Fernández · 2012
[ANGLES] Driver assistance systems are an important contribution to increase road safety. One aspect of the Advanced Driver Assistance Systems, ADAS, is to recognize traffic signs on the road, due to the amount of information given by them. The number of signs and the differences between them is considerably high, and the ADAS of Continental AG can already detect and classify circular signs. By acknowledging those facts, this thesis focuses on the signs with triangular shape. The following work presents two algorithms, named Loy-Barnes and the Fast Hough Transform. These algorithms are based on the so-called evidence accumulation method and the requirements to detect triangles efficiently. Both algorithms are analyzed, implemented, tested and compared in this thesis. The Loy-Barnes algorithm has proved to have a high detection rate, but also high runtime. The comparisons of this work should lead to the conclusion that the Fast Hough Transform implementation can also provide a high performance while achieving better runtime results.