摘要
随着我们对外景图像的质量要求越来越高,对有雾图像的去雾显得尤为重要。基于暗通道先验信息,在对大量有雾图像的雾气浓度分布分析后,总结出雾气分布与暗通道图像的关系,对暗通道图像进行三维曲面拟合估计图像景深。该算法兼顾去雾效果和处理效率,使得在原暗通道算法的基础上实时性得到大幅提升。
Removing haze from a single image is very important as the demand is being more and more higher for the quality of outdoor images. Based on dark channel prior, summaries up the relationship between dark channel and distribution of haze after distribution of haze analysis for a large number of images, three-dimensional surface fitting image depth of field for dark channel. The algorithm balanced the removing haze against the processing efficiency, compared with the primitive dark channel algorithm, real time can be a great improvement.
出处
《现代计算机》
2014年第1期26-30,34,共6页
Modern Computer
