计算机技术与发展Issue(2)：6-11,6.DOI:10.3969/j.issn.1673-629X.2015.02.002
改进的基于暗原色先验的图像去雾算法
An Improved Algorithm of Haze Removal Based on Dark Channel Prior
摘要
Abstract
Dark channel prior often gives a smaller value in transmission estimation,and doesn' t be suitable for the sky region,which leads to a dim image with cross-color sky regions after haze removal. To solve these problems,a new algorithm of haze removal based on sky region segmentation is proposed,which separates the sky region firstly,and use different methods to recover the scene radiance in sky and non-sky regions,overcoming the cross-color caused by failure of the prior rule. With a global White Patch adaptation in sky region and Gray World adaptation in non-sky region,it yields a brighter result,which is a good solution to solve the deficiency of the restored image' s dim. In addition,with estimating the atmospheric light only in sky regions,the method gives a more precise transmission and re-sults in more details in the final restored image. Experimental results show that this method can handle the cross-color of sky regions ef-fectively,and the restored result keeps sufficient brightness and contrast. It yields a better result with no much cost in time complexity.关键词
去雾/暗原色先验/引导滤波/天空分割/白平衡Key words
haze removal/dark channel prior/guided image filter/sky region segmentation/white balance分类
信息技术与安全科学引用本文复制引用
李坤,兰时勇,张建伟,李璐..改进的基于暗原色先验的图像去雾算法[J].计算机技术与发展,2015,(2):6-11,6.基金项目
国家“863”高技术发展计划项目(2013AA013802) （2013AA013802）
