矿井粉尘是危害煤矿人员和设备的重要原因之一，矿井粉尘浓度测量是煤炭安全管理的重要内容。由于井下环境复杂，传统矿用粉尘浓度传感器和粉尘浓度视频测量方法，无法消除光强和湿度对粉尘浓度测量的影响，且无法定位分布不均的粉尘区域，造成测量精度低，误报率高，测量面积有限等问题。针对以上问题，提出一种基于图像透射率的矿井粉尘浓度视频测量方法，并校正光强和湿度产生的测量误差，主要研究内容如下：

（1）矿井粉尘浓度视频测量方法研究。搭建模拟矿井粉尘浓度视频测量实验平台，采集不同光强和湿度下不同浓度粉尘图像；采用混合暗通道先验计算图像透射率t(x)，并通过皮尔逊相关系数法，分析不同光强下图像透射率t(x)与粉尘浓度值的相关性；利用引导滤波和全变差正则化优化图像透射率t(x)；围绕矿井粉尘浓度分布不均的问题，提出粉尘区域定位方法，结合图像透射率t(x)，实现矿井不同区域粉尘浓度值的计算。

（2）光强影响下矿井粉尘浓度视频测量误差校正方法研究。针对光强对粉尘浓度视频测量的影响，分析不同光强下图像透射率与粉尘浓度值，得出一定湿度下光强对粉尘浓度视频测量的影响规律；针对光强产生的测量误差，提出环境光监测方法，使光强监测准确度达92.4%；建立不同光强下粉尘浓度与图像透射率的数学模型校正误差，粉尘浓度测量准确度达93.36%，有效降低了矿井光强对粉尘浓度视频测量的影响。

（3）多因素影响下矿井粉尘浓度视频测量误差校正方法研究。针对湿度和光强对粉尘浓度视频测量的影响，通过对不同湿度和光强下图像透射率与粉尘浓度值的分析，得出不同光强下湿度对粉尘浓度视频测量的影响规律；针对湿度和光强产生的测量误差，提出基于Adamw双并联隐藏层BP网络校正误差，粉尘浓度测量准确度达95.71%，有效降低了矿井湿度和光强对粉尘浓度视频测量的影响。

本研究提出的粉尘浓度视频测量方法可有效降低矿井湿度和光强对粉尘浓度测量的影响，并准确计算不同区域粉尘浓度值。该方法为矿井粉尘浓度视频测量提供了技术支撑，对矿井粉尘浓度精准测量预警具有重要的现实意义。

Mine dust is one of the important reasons that harm personnel and equipment in coal mines. The measurement of mine dust concentration is an important content of coal safety management. Due to the complex underground environment, the traditional mining dust concentration sensor and dust concentration video measurement method can not eliminate the influence of light intensity and humidity on dust concentration measurement, and can not locate the uneven distribution of dust areas, resulting in low measurement accuracy, high false positive rate, limited measurement area and other problems. To solve the above problems, a video measurement method of mine dust concentration based on image transmittance is proposed, and the measurement errors caused by light intensity and humidity are corrected. The main research contents are as follows:

(1) Study on video measurement method of mine dust concentration. The experimental platform of simulating mine dust concentration video measurement was built to collect dust images of different concentrations under different light intensity and humidity. The image transmittance t(x) was calculated using mixed dark channel prior, and the correlation between image transmittance t(x) and dust concentration was analyzed by Pearson correlation coefficient method. The image transmittance t(x) is optimized by guided filtering and total variation regularization. In order to solve the problem of uneven distribution of dust concentration in mine, a method of locating dust regions is proposed, which is combined with image transmittivity t(x) to calculate the dust concentration in different regions of mine.

(2) Study on error correction method for video measurement of mine dust concentration under the influence of light intensity. According to the influence of light intensity on the video measurement of dust concentration, the image transmittance and dust concentration value under different light intensity are analyzed, and the influence rule of light intensity on the video measurement of dust concentration under certain humidity is obtained. Aiming at the measurement error caused by light intensity, an ambient light monitoring method is proposed, and the accuracy of light intensity monitoring is up to 92.4%. The mathematical model correction error of dust concentration and image transmission under different light intensity is established, and the dust concentration measurement accuracy is up to 93.36%, which effectively reduces the influence of mine light intensity on dust concentration video measurement.

(3) Study on error correction methods for video measurement of mine dust concentration under the influence of multiple factors. In view of the influence of humidity and light intensity on the video measurement of dust concentration, the law of influence of humidity and light intensity on the video measurement of dust concentration is obtained through the analysis of image transmission and dust concentration under different humidity and light intensity. In view of the measurement errors caused by humidity and light intensity, the accuracy of dust concentration measurement is 95.71% based on Adamw double parallel hidden layer BP network correction error, which effectively reduces the influence of mine humidity and light intensity on dust concentration video measurement.

The video measurement method of dust concentration proposed in this study can effectively reduce the influence of mine humidity and light intensity on dust concentration measurement, and accurately calculate the dust concentration value in different areas. This method provides technical support for the video measurement of mine dust concentration and has important practical significance for the accurate measurement and early warning of mine dust concentration.