Computer Engineering and Applications ›› 2018, Vol. 54 ›› Issue (12): 182-186.DOI: 10.3778/j.issn.1002-8331.1702-0123

Previous Articles     Next Articles

Image dehazing algorithm by wavelet fusion based on color attenuation prior

ZHANG Min, ZHANG Yifan, WANG Yuanyu   

  1. School of Computer Science and Technology, Taiyuan University of Technology, Jinzhong, Shanxi 030600, China
  • Online:2018-06-15 Published:2018-07-03


张  敏,张一凡,王园宇   

  1. 太原理工大学 计算机科学与技术学院,山西 晋中 030600

Abstract: An improved single image dehazing?algorithm using color attenuation prior and wavelet fusion is proposed in this paper. Firstly, according to color attenuation prior, a linear model of relationship between transmission and brightness and saturation is built, from which the rough transmission map can be calculated. Secondly, the reverse information of gray scale form of the hazy image is extracted as the detail supplement of the transmission map. Finally, the refined transmission map is obtained by fusing the rough transmission map and the reverse information of gray scale form of the haze image with wavelet algorithm. This method avoids the problem of manual picked parameters, improves the accuracy of transmission map by combining features of original haze images with highly-automatized. Experiments show that the proposed method is robust and restore images with good quality.

Key words: image dehazing, color attenuation prior, linear model, wavelet fusion

摘要: 针对图像去雾问题,在颜色衰减先验基础上提出了一种基于小波融合的单幅图像去雾方法。首先,通过颜色衰减先验假设建立了透射率关于图像亮度、饱和度的线性模型,估计出图像的粗略透射率信息。其次,提取雾图像灰度图的细节信息作为透射率的细节补充。最后,采用小波变换将两者进行融合,得到准确率高的透射率,进而恢复出清晰图像。该方法避免了大气散射系数的人工选择,自动化程度高。并且结合了原图像的特性,提高了透射率的准确性。实验表明该方法泛化效果好,恢复出的图像彩色自然。

关键词: 图像去雾, 颜色衰减先验, 线性模型, 小波融合