赵猛,任志浩,褚海峰,等. 基于大气散射模型的采煤工作面尘雾图像清晰化算法[J]. 煤炭学报,2023,48(8):3312−3322. DOI: 10.13225/j.cnki.jccs.2022.1131
引用本文: 赵猛,任志浩,褚海峰,等. 基于大气散射模型的采煤工作面尘雾图像清晰化算法[J]. 煤炭学报,2023,48(8):3312−3322. DOI: 10.13225/j.cnki.jccs.2022.1131
ZHAO Meng,REN Zhihao,CHU Haifeng,et al. Dust and fog image-sharpening algorithm based on atmospheric scattering model in coal face[J]. Journal of China Coal Society,2023,48(8):3312−3322. DOI: 10.13225/j.cnki.jccs.2022.1131
Citation: ZHAO Meng,REN Zhihao,CHU Haifeng,et al. Dust and fog image-sharpening algorithm based on atmospheric scattering model in coal face[J]. Journal of China Coal Society,2023,48(8):3312−3322. DOI: 10.13225/j.cnki.jccs.2022.1131

基于大气散射模型的采煤工作面尘雾图像清晰化算法

Dust and fog image-sharpening algorithm based on atmospheric scattering model in coal face

  • 摘要: 煤炭开采过程中产生了大量粉尘、水雾等悬浮颗粒,导致图像出现低照度、高尘雾且分布不均匀等降质现象,现有的图像清晰化方法处理效果不理想。提出一种基于大气散射模型的采煤工作面尘雾图像清晰化算法,解决了复杂采煤环境下模型参数(环境光值和透射率)估计不准问题。主要包括3个部分:根据采煤图像尘雾浓度分割图像区域;估计各区域的环境光值与透射率;融合区域参数后,基于大气散射模型恢复清晰图像。首先,通过分析采煤工作面尘雾分布特征,依据图像通道差异、亮度等信息,将采煤工作面尘雾区域分割为浓雾区域和非浓雾区域。然后,采用Max-RGB方法估计出尘雾图像初始光照图,为更好的保留尘雾图像结构信息和边缘信息,对初始光照图进行精细化处理,并利用精细化光照图分别计算出两区域的环境光值;在浓雾区域,依据尘雾浓度及分布特点,采用优化的颜色衰减模型,估计出浓雾区域的透射率;在非浓雾区域,利用暗通道先验以及该区域环境光矩阵,计算出非浓雾区域透射率。最后,对不同区域的环境光矩阵及透射率矩阵进行Alpha融合,并利用引导滤波在保留图像边缘信息的同时对融合过程产生的噪声进行抑制,得到全局的环境光和透射率,代入大气散射模型恢复低照度尘雾图像。为验证本文提出算法的有效性,选取了具有代表性的算法进行对比,试验表明提出算法能有效降低图像中雾的浓度,改善图像照度,保留图像边缘信息,总体性能更佳。

     

    Abstract: A large number of suspended particles such as dust and water fog is produced in the process of coal mining, which leads to the degradation of images such as low illumination, high dust concentration, and uneven fog distribution. The existing image-sharpening methods are not ideal. An algorithm based on atmospheric scattering model is proposed to restore the dust and fog image in coal face, which solves the problem of inaccurate estimation of atmospheric scattering model parameters in a complex coal mining environment. It mainly includes three parts: Segmenting the image according to the dust and fog concentration of the coal mining image; Estimating the ambient light value and transmittance of each region; After fusing the regional parameters, the clear image is restored based on atmospheric scattering model. Firstly, by analyzing the dust and fog distribution characteristics in coal face, according to the information of image channel difference and brightness, the dust and fog image of coal face is divided into dense fog region and non-dense fog region. Then, the Max-RGB method is used to estimate the initial illumination map of the dust and fog image. To better retain the structure information and edge information of the dust and fog image, the initial illumination map is refined, and the ambient light values of the two regions are calculated by using the refined illumination map; In the dense fog region, according to the dust and fog concentration and distribution characteristics, the transmittance is estimated by using the optimized color attenuation model; In the non-dense fog region, the transmittance is calculated by using the dark channel prior and the ambient light matrix of the region. Finally, the ambient light matrix and transmittance matrix in different regions are alpha fused, and the guidance filter is used to suppress the noise generated in the fusion process while retaining the image edge information, to obtain the global ambient light and transmittance values, which are substituted into the atmospheric scattering model to restore the low illumination of dust and fog image. In order to verify the effectiveness of the proposed algorithm, two representative algorithms are selected for comparison. Experiments show that the proposed algorithm can effectively reduce the dust and fog concentration in the image, improve the image illumination, retain the image edge information, and the overall performance is better.

     

/

返回文章
返回