|本期目录/Table of Contents|

[1]贺安民,孔旭辉,郑丽娜,等.基于图像机器学习的粉尘浓度估算方法*[J].中国安全生产科学技术,2025,21(1):79-86.[doi:10.11731/j.issn.1673-193x.2025.01.010]
　HE Anmin,KONG Xuhui,ZHENG Lina,et al.Estimation method of dust concentration based on image machine learning[J].JOURNAL OF SAFETY SCIENCE AND TECHNOLOGY,2025,21(1):79-86.[doi:10.11731/j.issn.1673-193x.2025.01.010]
点击复制

基于图像机器学习的粉尘浓度估算方法*
分享到：

《中国安全生产科学技术》[ISSN:1673-193X/CN:11-5335/TB]

卷:
21
期数:
2025年1期
页码:
79-86
栏目:
职业安全卫生管理与技术
出版日期:
2025-01-30

文章信息/Info

Title:
Estimation method of dust concentration based on image machine learning
文章编号:
1673-193X(2025)-01-0079-08
作者:
贺安民; 孔旭辉; 郑丽娜; 冯温婷; 李睿钢
（1.国能神东煤炭集团有限责任公司，陕西 榆林 719300；
2.中国矿业大学 安全学院，江苏 徐州 221116)
Author(s):
HE Anmin;  KONG Xuhui;  ZHENG Li’na;  FENG Wenting;  LI Ruigang
(1.Shendong Coal Co.,Ltd.,China Energy Coal Group,Yulin Shaanxi 719300,China;
2.School of Safety Engineering,China University of Mining and Technology,Xuzhou Jiangsu 221116,China)
关键词:
粉尘浓度检测; 图像法; 光散射; 图像灰度; 支持向量机
Keywords:
dust concentration detection;  image method;  light scattering;  image gray scale;  support vector machine
分类号:
X964
DOI:
10.11731/j.issn.1673-193x.2025.01.010
文献标志码:
A
摘要:
为了应对燃煤电厂超低排放标准的推行后，粉尘颗粒物排放特性向低浓度、小粒径方向发展的趋势，提出一种基于粉尘颗粒散射图像的浓度测量方法，以粉尘的侧向散射图像为研究对象，通过极度梯度提升树（eXtreme Gradient Boosting，XGBoost）、支持向量机回归（support vector regression，SVR）和随机森林（random forest，RF）算法，建立灰度值与粉尘浓度值之间的关系，从而实现粉尘浓度的检测。研究结果表明：基于粉尘散射图像建立的多变量输入XGBoost模型在对添加信号噪声前后的粉尘浓度预测中表现出色，拟合得到的R2接近于1，校正RMSE值在0.019 7～0.215 6之间，预测R2值在0.965 6～0.999 8的范围内，预测RMSE值在0.018 1～2.932 8之间，表明XGBoost模型具有较高的准确度以及稳定性。研究结果可为特定空间内某断面粉尘平均浓度的精确检测提供可靠依据，提高检测的准确性和便捷性。
Abstract:
In order to cope with the development trend of dust particle emission characteristics towards low concentration and small particle size after the implementation of ultra-low emission standards for coal-fired power plants,a concentration measurement method based on the scattering images of dust particles was proposed.The lateral scattering image of dust was taken as the research object,and the relationship between gray value and dust concentration value was established by eXtreme Gradient Boosting (XGBoost),support vector regression (SVR) and random forest (RF) algorithm,so as to realize the detection of dust concentration.The results show that the multivariable input XGBoost model established based on dust scattering images performs well in predicting the dust concentration before and after adding signal noise.The fitted R2 is close to 1,the calibration RMSE value is between 0.019 7 and 0.215 6,the predicted R2 value is between 0.965 6 and 0.999 8,and the predicted RMSE value is between 0.018 1 and 2.932 8,which indicates that the XGBoost model has high accuracy and stability.The research results can provide a reliable basis for the accurate detection of the average dust concentration in a certain section of specific space,improving the accuracy and convenience of detection.

参考文献/References:

［1］戴彦德.我国可持续发展中的能源问题［J］.节能,2003(2):3-5. DAI Yande.Energy issues in China’s sustainable development［J］.Energy Conservation,2003(2):3-5.
［2］陈栋.颗粒特性对光散射法在线测量颗粒物质量浓度的影响研究［D］.武汉:华中科技大学,2019.
［3］XU M,YU D,YAO H,et al.Coal combustion-generated aerosols:Formation and properties［J］.Proceedings of the Combustion Institute,2011,33(1):1681-1697.
［4］POPE C A,BURNETT R T,THUN M J,et al.Lung cancer,cardiopulmonary mortality,and long-term exposure to fine particulate air pollution［J］.JAMA,2002,287(9):1132.
［5］SEINFELD J H,PANDIS S N.Atmospheric chemistry and physics:from air pollution to climate change［M］.New York:Wiley,1998.
［6］中华人民共和国卫生部.工作场所空气中粉尘测定 第1部分:总粉尘浓度:GBZ/T 192.1—2007［S］.北京:人民卫生出版社,2007.
［7］中华人民共和国卫生部.工作场所空气中粉尘测定 第2部分:呼吸性粉尘浓度:GBZ/T 192.2—2007［S］.北京:人民卫生出版社,2007.
［8］陈清华,许曾生,王小润,等.基于滤膜称重法的自动化粉尘质量浓度检测装置的研究 ［J］.煤炭学报,2024,49(7):2997-3006. CHEN Qinghua,XU Zengsheng,WANG Xiaorun,et al.Design of automatic dust mass concentration detection device based on membrane weighing method［J］.Journal of China Coal Society:2024,49(7):2997-3006.
［9］魏海天,王杰,陈述斌.基于数据融合的光散射法与β射线粉尘监测技术研究 ［J/OL］.矿业安全与环保,1-7［2024-11-09］.http://kns.cnki.net/kcms/detail/50.1062.td.20240415.1651.002.html.
［10］赵政.基于光散射多源耦合的粉尘浓度检测技术［J］.煤矿安全,2023,54(8):195-201. ZHAO Zheng.Dust concentration detection technology based on light scattering multi-source coupling［J］.Safety in Coal Mines,2023,54(8):195-201.
［11］陈建阁,李德文,许江,等.基于光散射法无动力粉尘质量浓度检测技术［J］.煤炭学报,2023,48(增刊1):149-158. CHEN Jian’ge,LI Dewen,XU Jiang,et al.Detection technology of unpowered dust concentration based on light scattering method［J］.Journal of China Coal Society,2023,48(Supplement 1):149-158.
［12］余晓轲,柯丽华,张光权,等.基于图像技术的爆区粉尘浓度测量方法研究［J］.矿业研究与开发,2024,44(4):168-174. YU Xiaoke,KE Lihua,ZHANG Guangquan,et al.Measurement method of dust concentration in blasting area based on image technology［J］.Mining Research and Development,2024,44(4):168-174.
［13］SABA T,REHMAN A,AL-DHELAAN A,et al.Evaluation of Current Documents Image Denoising Techniques:A Compar-ative Study［J］.Applied Artificial Intelligence,2014,28(9):879-887.
［14］褚江,陈强,杨曦晨.全参考图像质量评价综述［J］.计算机应用研究,2014,31(1):13-22. CHU Jiang,CHEN Qiang,YANG Xichen.Review on full reference image quality assessment algorithms［J］.Application Research of Computers,2014,31(1):13-22.
［15］SANKUR B.Statistical evaluation of image quality measures［J］.Journal of Electronic Imaging,2002,11(2):206.
［16］张宸瑜.烟尘浓度光散射测量技术研究［D］.南京：东南大学,2015.
［17］邢键.基于光后向散射法的烟尘浓度测量技术研究［D］.哈尔滨：哈尔滨工业大学,2010.
［18］方庆,谢岸煌,谢仁礼.一种自适应反锐化掩膜清晰度增强算法［J］.电视技术,2022,46(8):15-19. FANG Qing,XIE Anhuang,XIE Renli.An adaptive unsharp masking sharpness enhancement algorithm［J］.Video Engineering,2022,46(8):15-19.
［19］HEIL T,TATLOCK G J.Noise reduction in CCD measurements by improving the quality of dark-reference images［J］.Microscopy,2018,67(suppl_1):i123-i132.
［20］韩思奇,王蕾.图像分割的阈值法综述［J］.系统工程与电子技术,2002(6):91-94,102. HAN Siqi,WANG Lei.A survey of thresholding methods for image segmentation［J］.Systems Engineering and Electronics,2002(6):91-94,102.
［21］SMITH J P,BARON P A,MURDOCK D J.Response characteristics of scattered light aerosol sensors used for control monitoring［J］.American Industrial Hygiene Association Journal,1987,48(3):219-229.
［22］GRNER P,BEMER D,FABRIS J F.Photometer measurement of polydisperse aerosols［J］.Journal of Aerosol Science,1995,26(8):1281-1302.
［23］丁世飞,齐丙娟,谭红艳.支持向量机理论与算法研究综述［J］.电子科技大学学报,2011,40(1):2-10. DING Shifei,QI Bingjuan,TAN Hongyan.An overview on theory and algorithm of support vector machines［J］.Journal of University of Electronic Science and Technology of China,2011,40(1):2-10.
［24］潘梦瑶,任瑛,王思源,等.基于梯度提升算法和SHAP的石家庄PM2.5和臭氧浓度预测及影响因素分析［J］.环境科学学报,2024,44(7):402-409. PAN Mengyao,REN Ying,WANG Siyuan,et al.Prediction of PM2.5 and ozone concentration in Shijiazhuang and analysis of influencing factors based on gradient boosting algorithm and SHAP.Acta Scientiae Circumstantiae:2024,44(7):402-409.
［25］赵耀忠,严俊龙,任吉凯,等.基于机器学习的露天煤矿粉尘浓度预测［J］.煤炭工程,2022,54(增刊1):157-161. ZHAO Yaozhong,YAN Junlong,REN Jikai,et al.Dust concentration prediction of open-pit coal mine based on machine learning［J］.Coal Engineering,2022,54(Supplement 1):157-161.

相似文献/References:

备注/Memo

备注/Memo:
收稿日期： 2024-05-20
* 基金项目： 国家重点研发计划项目(2023YFC3010602)；中央高校基本科研业务费专项资金项目(2021YCPY0107)
作者简介： 贺安民，博士，高级工程师，主要研究方向为采掘技术、一通三防。
通信作者： 郑丽娜，博士，教授，主要研究方向为工作场所粉尘监测、拉曼光谱定量分析。
更新日期/Last Update: 2025-01-26