计算机工程与应用 ›› 2017, Vol. 53 ›› Issue (8): 209-215.DOI: 10.3778/j.issn.1002-8331.1509-0272

• 图形图像处理 • 上一篇    下一篇

基于颜色饱和度的快速图像去雾研究

刘万军,赵庆国   

  1. 辽宁工程技术大学 软件学院,辽宁 葫芦岛 125105
  • 出版日期:2017-04-15 发布日期:2017-04-28

Fast image dehaze study based on color saturation

LIU Wanjun, ZHAO Qingguo   

  1. Software Engineering Institute, Liaoning Technical University, Huludao, Liaoning 125105, China
  • Online:2017-04-15 Published:2017-04-28

摘要: 在雾霾天气条件下,室外场景图像往往会产生严重退化,造成图像对比度的下降和颜色的衰减。为得到清晰无雾的图像,从单色大气散射模型出发,提出一种基于颜色饱和度的快速图像去雾算法(FIDS)。首先,大气光亮度[A]通过四叉树方法进行估计,接着采用饱和度运算对透射率进行估计,并对其修正和平滑保边得到较精确的大气透射率,最后基于大气散射模型获取最终的去雾图像。对比实验结果表明该算法能提高图像的清晰度和运算效率,很好地恢复图像的颜色和对比度。

关键词: 图像去雾, 大气散射模型, 四叉树, 饱和度

Abstract: Under the condition of the haze weather, the outdoor scene images tend to have a serious degradation, resulting in the decrease of the image contrast and the color attenuation. In order to obtain clear haze-free image, this paper puts forward a fast color saturation based image to fog algorithm(FIDS), using quad tree to estimate the atmospheric brightness [A], using saturation arithmetic to estimate the transmissivity function and correction and edge preserving smoothing operation, and the atmospheric scattering model obtains the final fog image from monochrome atmospheric scattering model. Experimental results show that the proposed algorithm can recover the contrast and color of the image, improve the visual definition and operation efficiency of the image.

Key words: image defog, atmospheric scattering model, quadtree, saturation