电子学报 ›› 2015, Vol. 43 ›› Issue (7): 1437-1443.DOI: 10.3969/j.issn.0372-2112.2015.07.029

• 科研通信 • 上一篇    下一篇

一种基于暗通道先验的快速图像去雾算法

张登银, 鞠铭烨, 王雪梅   

  1. 南京邮电大学物联网学院, 江苏南京 210003
  • 收稿日期:2013-11-26 修回日期:2014-04-25 出版日期:2015-07-25
    • 作者简介:
    • 张登银 男,1964年11月出生,江苏靖江人,研究员、博士生导师.1986年、1989年和2004年在南京邮电大学分别获工学学士、硕士和博士学位,2008年瑞典UMEA大学数字媒体实验室博士后出站.主要研究领域是下一代信息网络、信号与信息处理、网络信息安全. E-mail:wangxm@njupt.edu.cn;鞠铭烨 男,1989年1月出生,江苏靖江人,博士生.2013年毕业于天津工程师范学院,获工学硕士学位,研究方向为数字图像处理、视频压缩.
    • 基金资助:
    • 国家自然科学基金 (No.61071093); 国家863高技术研究发展计划 (No.2010AA701202); 瑞典-亚洲国际合作项目 (No.348-2008-6212); 江苏省重大科技支撑计划 (No.BE2012849); 江苏省普通高校研究生科研创新计划 (No.CXZZ13_0476); 江苏省产学研前瞻性联合研究项目 (No.BY2014014)

A Fast Image Daze Removal Algorithm Using Dark Channel Prior

ZHANG Deng-yin, JU Ming-ye, WANG Xue-mei   

  1. School of Internet of Things, Nanjing University of Posts and Telecommunications, Nanjing, Jiangsu 210003, China
  • Received:2013-11-26 Revised:2014-04-25 Online:2015-07-25 Published:2015-07-25
    • Supported by:
    • National Natural Science Foundation of China (No.61071093); National High Technology Research and Development Program of China  (863 Program) (No.2010AA701202); Swedish-Asia International Cooperation Project (No.348-2008-6212); Major Science and Technology Support Project of Jiangsu Province (No.BE2012849); Graduate Research Innovation Program of univerities in Jiangsu Province (No.CXZZ13_0476); Prospective Project of Industry,  University and Research of Jiangsu Province (No.BY2014014)

摘要:

目前针对单幅图像去雾处理,暗通道先验算法具有较好的效果,但处理时间长,需要很大的储存资源与计算资源.本文在暗通道先验算法基础上提出一种快速去雾算法,首先用"边缘替代法"代替原算法中的抠图处理,显著降低了计算复杂度;然后针对明亮区域暗通道失效情况,提出了一种基于双阀值的明亮区域识别方法和透射率修正机制,提高了暗通道先验的适用范围;最后对去雾图像再采用非线性对比度拉伸进行增强处理,改善了去雾图像的视觉效果.实验表明:本文算法不仅极大地降低了计算复杂度,而且去雾能力优于原算法.

关键词: 图像去雾, 暗通道先验, 边缘替代, 透射率

Abstract:

While the dark channel prior algorithm works well for single image haze removal,it has to take long time to proces and require expensive computational overhead and huge memory resource.In this paper,we propose a fast daze removal algorithm based on dark channel prior knowledge.In the proposed approach,edge substitution method is developed to take place of the original soft-matting operator,which will substantially reduce the computational complexity.To deal with the invalid case of the dark channel,we exploit a scheme to identify bright area based on dual threshold and develop a way to correct transmission rate,which enables the dark channel prior to be more applicable.Additionally,enhanced processing is introduced by adopting nonlinear contrast stretching,thus improving the visual effect of defogging images.Several numerical experiments demonstrate that the proposed approach outperforms the original algorithm in daze removal effect at a much lower implementation cost.

Key words: image daze removal, dark channel prior, edge substitution, transmission

中图分类号: