期刊名称：International Journal of Signal Processing, Image Processing and Pattern Recognition
印刷版ISSN：2005-4254
出版年度：2015
卷号：8
期号：7
页码：171-178
DOI：10.14257/ijsip.2015.8.7.16
出版社：SERSC
摘要：To solve the problem of image definition reduction caused by haze in intelligent transportation monitoring, an algorithm of fast recovering for image degraded by haze based on dark primary colors prior. Firstly, we estimate atmospheric light and transmission rate using global dark colors prior based on the shortage of existing dark colors prior. Secondly, according to the atmospheric physics model, image degraded by haze is recovered. Finally, brightness of image is compensated by using Retinex theory. Experimental results show that the proposed algorithm can increase image definition better and the calculation time is 30 frames per second
关键词：intelligent transportation; faze; definition; image recovery; dark colors ; prior
