Authors: Hongjun Song; Yangzhou Chen; Yuanyuan Gao
Addresses: Beijing University of Technology, Beijing, BJ, 100124, China ' Beijing University of Technology, Beijing, BJ, 100124, China ' Zhejiang A&F University, Hangzhou, HZ, 311300 Zhejiang, China
Abstract: A real time traffic meteorological visibility distance evaluation algorithm in foggy weather by using dark channel prior and lane detection methodology is proposed in this paper. In foggy image, dark channel prior directly provides accurate transmission estimation. A novel lane detection algorithm which is called variable box search (VBS) is proposed in this paper. This novel algorithm only needs little running time and could maintain real time procession. Background generating and updating method which is called Gaussian mixture model (GMM) will be used to get clear background image. Two endpoints of one traffic lane are marked and saved; these data will be served for traffic scene distance calculation. Extinction coefficient k could be calculated by these two end points transmission division based on the monocular model and dark channel prior. Finally, the meteorological visibility will be according to definition from International Commission on Illumination. According to the traditional fog sorting methodology, we fulfil fog category method by our algorithm based on the extinction coefficient value. Experimental data are taken from the actual traffic scene and network data. Experimental results verify the effectiveness of this proposed algorithm.
Keywords: visibility distance evaluation; monocular model; dark channel prior; lane detection; variable box search; foggy weather; fog; Gaussian mixture model; GMM; traffic distances; vehicle visibility; extinction coefficient value.
International Journal of Computational Science and Engineering, 2015 Vol.10 No.4, pp.375 - 386
Received: 08 Dec 2011
Accepted: 18 Jul 2012
Published online: 05 Aug 2015 *