An Adaptive Approach to Detect Warning Traffic Signs using SOM and Windowed Hough Transform

Hasan Fleyeh, Rubel Biswas, and Nizam U. Bhuiyan


Traffic signs, Sign Detection, Hough Transform, Color segmentation, Self Organizing Maps


Warning traffic signs represent an important group of traffic signs which indicate danger for road users. Detecting this group in good time may be helpful to avoid many fatal accidents. This paper presents a new approach to detecting warning traffic signs which is based on color segmentation using Self Organizing Maps and windowed Hough Transform. The proposed system is a standalone and adaptive which means that it works without any kind of training. This is due to the fact that color segmentation using SOM employs 1% of the image under investigation for the training and Hough Transform is invoked to detect the shape of this group of traffic signs. Experiments conducted to check the robustness of this approach indicated that 95.6% of the traffic signs invoked for this test were successfully detected. This test was carried out under a wide range of environmental conditions and in different European countries.

Important Links:

Go Back