结合物理模型和Retinex的图像去雾算法

Image Fog Removal Based on Physics Model and Retinex Algorithm

在雾霾等天气下拍摄到的图像,存在对比度降低、颜色失真、模糊不清等严重的退化现象.针对这一问题,提出一种新的单幅图像去雾算法.首先运用物理模型和Retinex算法相结合估算透射率,提高了运算效率,有效避免了晕光现象和色彩失真.其次用自适应权值融合计算大气光,分割出图像的亮区域和暗区域自适应估算融合权值,使获得的大气光值更加合理精确.最后增加可控参数限制透射率过低,反解物理模型复原图像.实验分析表明本文算法简单有效,复原出的图像清晰自然,达到了很好的去雾效果.

Images taken in foggy weather exist serious degraded phenomena such as low contrast, color distortion and obscure. A new haze removal algorithm from a single image is proposed. First, physical model and retinex algorithm are combined to estimate trans- mission which improves computation efficiency and prevents the halo of light and color distortion. Then, an adaptive weighted fusion method is proposed to calculate atmospheric light. The fusion weights are estimated via image dark and bright segmentation. The ob- tained atmosphere light value is more reasonable and accurate. Finally, a parameter is added for image recovery to limit the low trans- mission, and a clear image is recovered by inverse transformation of physical model. Experiments show that the algorithm is simple and effective, and recovered images are clear and natural, achieving a good defogging effect.