摘要
针对基于暗原色先验处理后,由于图像偏暗以及天空区域出现颜色偏移、失真等现象,从而造成去雾不彻底的问题,在暗原色先验理论的基础上,提出了一种基于引导滤波和变差函数的图像去雾算法。首先,使用两次引导滤波对暗原色先验模型的透射率进行优化,然后利用变差函数来选取合适的阈值,从而选取准确的大气光值,最后将得到的参数代入去雾物理模型中完成去雾处理。实验结果表明,该算法可以对大气光进行准确估计,有效避免了天空区域的影响。通过不同的算法对不同的室外采集的雾天图像的对比效果可知,该算法可以较好地处理带有光的雾天图像,恢复出来的图像具有更好的细节保持性且更加清晰,视觉效果更加。
In order to solve the problem of incomplete defogging caused by dark image and color deviation and distortion in the sky area after processing based on dark channel prior,based on the theory of dark channel prior,we propose an image dehazing algorithm based on guided filtering and variogram. First,the transmission of the dark channel prior model is optimized by two guided filters,and then the variogram is used to select the appropriate threshold to select the accurate atmospheric light value. Finally,the obtained parameters are substituted into the dehazing physical model for defogging. The experiment shows that the proposed algorithm can accurately estimate the atmospheric light and effectively avoid the influence of the sky region. Through different algorithms to compare the contrast effects of different outdoor foggy images,the proposed algorithm can better deal with foggy images with light. The restored images have better detail retention and clearer visual effects.
作者
章星晨
孙刘杰
ZHANG Xing-chen;SUN Liu-jie(University of Shanghai for Science and Technology,Shanghai 200093,China)
出处
《计算机技术与发展》
2019年第6期23-26,共4页
Computer Technology and Development
基金
上海市科学技术计划项目(18060502500)
上海理工大学科技发展项目(16KJFZ017)
关键词
图像去雾
暗原色先验
透射率
大气光
引导滤波
变差函数
image defogging
dark channel prior
transmission map
high lightregions
guided filtering
variogram
