计算机与现代化

• 图像处理 • 上一篇    下一篇

 基于暗原色先验的海边雾天图像去雾

  

  1. 青岛大学自动化与电气工程学院,山东青岛266071
  • 收稿日期:2016-03-17 出版日期:2016-07-21 发布日期:2016-07-22
  • 作者简介: 霍海蒙(1988-),女,山东菏泽人,青岛大学自动化与电气工程学院硕士研究生,研究方向:数字图像处理; 徐世许(1963-),男,教授,博士,研究方向:机器视觉,人工智 能,图像处理,导航,制导与控制; 王汉萍(1982-),女,副教授,研究方向:人工智能,数字图像处理。

 Beach Foggy Images Haze Removal Based on Dark Channel Prior

  1. College of Automation and Electrical Engineering, Qingdao University, Qingdao 266071, China
  • Received:2016-03-17 Online:2016-07-21 Published:2016-07-22

摘要:

 在雾、霾等天气的影响下,获取的海边图像严重失真降质。为了更好地恢复图像得到有用的信息,结合暗原色先验理论,提出K-means聚类算法分割天空区域的方法。首先针对海边场景所
具有的特性利用K-means聚类算法分割天空区域,大气光强度的值是将非天空区域的亮度值与天空区域亮度值加权得到的;其次基于大气散射模型对其估计透射率,并对透射率进行改善;最后为了增加图
像的亮度,增强对比度,对复原图像进行色调重映射。通过大量实验表明,该算法针对海边图像有很好的去雾效果,与传统的去雾算法相比,该算法去雾效果更佳,更适合于处理海边雾天图像。

关键词:  , 暗原色先验, 大气散射模型, K均值聚类

Abstract:

 In the effects of fog, haze and other bad weather, the beach images acquired from camera have serious distorted and degraded. In order to get better picture
information from restored image, this paper proposed a K-means clustering algorithm combined with dark channel prior theory. Firstly, using K-means clustering algorithm to
divide sky region based on the characteristics of beach scene, the atmospheric light intensity value is obtained by non-sky area luminance values with the sky region luminance
values weighted. Secondly, the transmission is estimated and improved based on atmospheric scattering model. Finally, in order to increase the brightness of the image and
enhance the image contrast, this paper adopted tone remapping on the image. A large number of experiments show that the algorithm can get an excellent performance on image haze
removal for the beach foggy image and particularly do better than traditional defogging algorithm, which is more suitable for processing beach foggy image.

Key words:  dark channel prior, atmosphere scattering model, K-means clustering