An efficient framework for recognizing traffic lights in night traffic images
Bo Fan, Weiyao Lin, Xiaokang Yang · 2012
In this paper, we propose a new and efficient framework for detecting and recognizing traffic signal lights with the corresponding countdown counters in the night scenes. The proposed framework first compares the difference between the daytime and night situation of traffic lights in the image and locate the recognition candidates in the HLC color space. With the locating of the main location marks, the shooting angle direction could be corrected. Based on the frontal direction, the content of the traffic lights can be recognized with the template matching measure. Furthermore, in order to reduce computation, a complexity reducing method is proposed. Experimental results on various scenes demonstrate the effectiveness of our proposed framework.