Color-based detection of vehicle lights

Abstract

In this article, we present a first approach in the conception of a vision system dedicated to the detection of vehicles in reduced visibility conditions. It is based on a self-adaptive stereo vision extractor of 3D edges or obstacle and a color detection of vehicle lights. The detection of vehicle lights uses the L*a*b* color space. The vision system detects three kinds of vehicle lights: rear lights and rear-brake-lights; flashing and warning lights; reverse lights and headlights.