A moving objects based real-time defogging method for traffic monitoring videos
暂无分享,去创建一个
In this paper, a moving objects based real-time defogging method for traffic monitoring videos is proposed. Firstly, dark channel prior based image defogging method has been improved. Then, the proposed image defogging method is used for traffic monitoring video defogging. To improve the processing speed, the correlation between the adjacent frames of videos is exploited. The moving objects are detected using adjacent frame difference method. The frame content is divided into moving foreground and background. Afterwards, the foreground and background are processed with different defogging manners to reduce the computational complexity of defogging processing. Experimental results show that the proposed method can generate a good defogging effect which will facilitate the subsequent intelligent traffic analysis. Furthermore, the proposed method is fast enough to process the standard-definition videos at the speed of 26 frames per second on average.
[1] Tang Jin. Review and prospect of image dehazing techniques , 2010 .
[2] Xie Bin,et al. Video Defogging Algorithm Based on Fog Theory , 2011 .
[3] 孙伟 Sun Wei,et al. Fast single image fog removal based on atmospheric scattering model , 2013 .
[4] Mohinder Malhotra. Single Image Haze Removal Using Dark Channel Prior , 2016 .
[5] Qi Mei-bin,et al. Improved algorithm on image haze removal using dark channel prior , 2011 .