煤矿井下散射环境雾化图像的气室仿真方法

An air chamber simulation method for scattering ambient foggy images in underground coal mines

在粉尘环境中,井下监控图像会出现雾化退化现象,而在井下难以进行环境控制实验开展图像雾化的研究。因此,文章利用相对平均梯度定义图像雾化程度、分析了井下粉尘粒径构成、提出了井下粉尘配比方法、设计并搭建仿真气室模拟井下雾化图像。结果表明:利用平均梯度定义的雾化程度优于其他常用评价参数,其评价结果与人眼判断结果相符,而井下图像的雾化退化程度可以达到80%,此时图像的细节信息基本被雾化影响所掩盖;井下粉尘样本的测试结果显示:99%的井下粉尘粒径小于30μm,粒径小于10μm的粉尘的粒子数占比约为93%;利用仿真气室生成的雾化图像不仅可以涵盖井下图像雾化程度范围,相较于雾化算法生成的结果,具有更准确地模拟粉尘下的光散射扩散效果的优点。

In underground mining environment,surveillance images are subject to fogging degradation because of dust scattering and it is difficult to conduct environmentally controlled experiments in situ for fogging and defogging investigation.Therefore,we define the degree of image fogging using the relative average gradient,analyze the composition of the dust particle sizes in the mine,propose the method for compose experiment powder to simulate different mining dust,and design and build an air chamber to simulate the mining scattering environment.The results indicate that:Firstly,the degree of fogging defined by the average gradient is better than other commonly used evaluation parameters and its evaluation results are consistent with the human eye perception;the degree of fogging of underground mining images can reach 80%,when the image details merge in the foggy blurring effect.Secondly,with microscopic observation method,99%of underground mining dust has particle size less than 30μm and the proportion with the particle size less than 10μm is about 93%.Thirdly,the foggy image generated from the simulated air chamber not only covers the range of the image fogging degree of real mining surveillance images,but also shows the diffusion effect of light scattering,which is more consistent with real surveillance images compared with computer-based fogging algorithm,proving the advantage of the air chamber method.