Cao Pingping, Wang Xianchao, Li Linguo, Liu Mingjun, Wang Mengting
School of Computer and Information Engineering, Fuyang Normal University, Fuyang, 236041, China.
Anhui Engineering Research Center for Intelligent Computing and Information Innovation, Fuyang Normal University, Fuyang, 236041, China.
Sci Rep. 2025 Apr 2;15(1):11235. doi: 10.1038/s41598-025-95912-z.
Due to the lack of underground space and lighting in coal mines, coal mine images suffer from low contrast, poor clarity and uneven brightness, which severely obstacles the visual task achievement in underground coal mines. Since the coal mine dust image has a special black shift, the existing ground and underwater defogging methods cannot play a role in the coal mine dust image with the black shift. Therefore, this paper proposes a method of coal mine dust image defogging with a three-stream and three-channel color balance, which is specially used for the restoration of disturbed coal mine images. The method performs color balance on the image R, G, and B channels respectively to eliminate the color shift caused by the coal mine environment; then uses a quad-tree subdivision search algorithm and dark channel prior to obtain the atmospheric light and transmittance of the three-channel color balanced image, respectively; then proposes a weighting algorithm to realize transmittance fusion of three-stream coal mine images, and finally realizes coal mine dust image defogging according to the haze weather degradation model. Extensive experimental results on the ground, underwater, sand and dust images and real coal mine images show that our method outperforms state-of-the-art coal mine dust image defogging algorithms and has good generality.
由于煤矿井下空间和光照条件的限制，煤矿图像存在对比度低、清晰度差和亮度不均匀等问题，这严重阻碍了井下视觉任务的完成。由于煤矿粉尘图像存在特殊的黑移现象，现有的地面和水下图像去雾方法在具有黑移现象的煤矿粉尘图像上无法发挥作用。因此，本文提出了一种基于三流三通道颜色平衡的煤矿粉尘图像去雾方法，专门用于恢复受干扰的煤矿图像。该方法分别对图像的R、G、B通道进行颜色平衡，以消除煤矿环境引起的颜色偏移；然后利用四叉树细分搜索算法和暗通道先验分别获取三通道颜色平衡图像的大气光和透射率；接着提出一种加权算法实现三流煤矿图像的透射率融合，最后根据雾霾天气退化模型实现煤矿粉尘图像的去雾。在地面、水下、沙尘图像和真实煤矿图像上的大量实验结果表明，我们的方法优于现有最先进的煤矿粉尘图像去雾算法，具有良好的通用性。
