An adaptive segmentation algorithm for traffic lights image based on OTSU and fuzzy theory

In order to improve the accuracy of traffic lights segmentation in complex background, a novel adaptive threshold algorithm based on Otsu and fuzzy theory is proposed. Firstly, the fuzzy area was obtained by rough segmentation based on Otsu theory. Then, taking into account the characteristic of traffic lights images in complex background, the membership function about luminance component of image is constructed. Finally, along with defuzzification by the gravity method, the threshold which could segment image adaptively was obtained. The experimental results demonstrate that our algorithm provides more accurate traffic light segmentation than the existing state-of-the-art algorithm through qualitative and quantitative evaluations.