Intelligent headlight control using camera sensors

Abstract

This paper describes our recent work on intelligently controlling a vehicle's headlights using a forward-facing camera sensor. Specifically, we aim to automatically control its beam state (high beam or low beam) during a night-time drive based on the detection of oncoming/overtaking/leading traffics as well as urban areas from the videos captured by the camera. A three-level decision framework is proposed which includes various types of image and video content analysis, an SVM-based learning mechanism and a frame-level decision making mechanism. Both video and context information have been exploited to accomplish the task. Online test drives as well as offline evaluations on tens of videos have validated the robustness and effectiveness of the proposed system.Copyright 2009 ACM.