Research on Warning System of Dense foggy Days Based on Image Processing

Abstract

On dense foggy days, the visibility is low. In order to increase the driving safety of vehicles, this paper studies a set of image-based dense foggy day discrimination method and warning system, which provides early technical supports for the intelligent start and stop of fog lamps. Firstly, by analysing and comparing image characteristics of sunny and dense foggy days, finding out the obvious distinction between images on sunny days and dense foggy days: the dark channel features. And then using the support of vector machine to obtain the judgment model of sunny/dense foggy day images. Finally, using MATLAB GUI to build a set of dense foggy day warning system which is based on image processing. After testing, the system is able to effectively distinguish between sunny and dense foggy day images with certain accuracy.