Abstract: Abstract：Introduce a kind of improved algorithm which is used to process the images cranking with the condition of low visibility. We use the physical model of atmosphere to recovery the quality of images as the condition in the sunshine. This measure has been described in some papers before. The different between the algorithm here and the measure in other papers before is that we do not use the RGB colour space but HSI colour space, and when we seek for the “disappeared point” and divide up the area of sky the algorithm does not depand on human eyes but can finish with itself,that is it has self adaptation. The emulation by MATLAB shows that this algorithm has the greater effect and has a well practicality in image processing.

Key words: physical, image recovery, HSI colour space, self adaptation

摘要： 摘要：介绍了一种对低能见度下所拍摄图像进行处理的改进算法，通过建立大气的物理模型，对图象进行复原处理，使图像质量接近晴天下的情况。与以往文献不同之处在于，在物理模型中使用了HSI而非RGB彩色空间，在寻找消失点和分割天空区域时，不是依靠人眼的主观选择，而由本改进算法自行完成，有一定的自适应性。MATLAB仿真结果表明，该算法处理效果有了进一步的提高，有较强的应用价值。

关键词: 物理模型, 图像复原, HIS彩色空间, 自适应