您的位置：山东大学 -> 科技期刊社 -> 《山东大学学报（理学版）》

《山东大学学报(理学版)》 ›› 2024, Vol. 59 ›› Issue (4): 127-134.doi: 10.6040/j.issn.1671-9352.0.2023.376

•   • 上一篇    

基于DEM与“宽带结构”联合优化的XCH4遥感反演算法研究

王晨(),许德刚,达虹鞠,唐智和,栾辉,范海浩   

  1. 中国石油集团安全环保技术研究院有限公司, 石油石化污染物控制与处理国家重点实验室，北京 102206
  • 收稿日期:2023-07-31 出版日期:2024-04-20 发布日期:2024-04-12
  • 作者简介:王晨(1988—)，男，工程师，硕士，研究方向为碳足迹技术. E-mail：wangch666@cnpc.com.cn
  • 基金资助:
    中国石油天然气集团有限公司科技项目(2021DQ03-A3)

Research on XCH4 remote sensing inversion algorithm based on DEM and broadband structure joint optimization

Chen WANG(),Degang XU,Hongju DA,Zhihe TANG,Hui LUAN,Haihao FAN   

  1. State Key Laboratory of Petroleum Pollution Control, CNPC Research Institute of Safety and Environment Technology, Beijing 102206, China
  • Received:2023-07-31 Online:2024-04-20 Published:2024-04-12

RichHTML

0

PDF (PC)

45

摘要：

加权修正的差分光学吸收光谱法(weighting function modified differential optical absorption spectroscopy，WFM-DOAS)是用于甲烷平均干空气摩尔分数(XCH4)遥感反演的经典算法，其关键技术之一是分离“宽带吸收”与“窄带吸收”光谱结构；同时，数字高程模型(digital elevation model, DEM)对XCH4的反演有重要影响。目前已有的甲烷反演产品主要使用多项式进行宽带结构拟合，多项式阶数的选择标准不明确、对宽带结构的拟合不够精确，使用的DEM精度无法满足局部地区高精度反演要求。本文选取瓦里关大气本底基准观象台所在的青藏高原区域为研究区，使用更高精度的数字高程模型(global 30 m digital elevation model, GLO-30)并用全连接神经网络代替低阶多项式进行“宽带结构”拟合，进一步地，在传统的全连接神经网络的基础上加入了“跳连”结构，并使用dropout策略对网络进行优化。将实验结果与使用The Global Multi-resolution Terrain Elevation Data 2010(GMTED2010)和低阶多项式拟合方法下反演的XCH4进行数据对比。结果显示，改进后的全连接神经网络可以更好地拟合宽带光谱结构，同时联合更高精度的DEM可以提高XCH4的反演精度，相关系数最高提高到0.92。所使用的联合优化方法可以用于油气田产区的XCH4的遥感反演，从而更好地服务于油气田产区甲烷异常排放排查等。

关键词: 加权修正的差分光学吸收光谱法, XCH4, DEM, 光谱宽带结构, 人工神经网络, 卫星遥感

Abstract:

A well-known approach for remote sensing inversion of the mean dry air mole fraction of methane (XCH4) is weighted function modified differential optical absorption spectroscopy (WFM-DOAS). The ability to distinguish between the spectral structures of "broadband absorption" and "narrowband absorption" is one of its primary technologies. In the meantime, the inversion of XCH4 is greatly influenced by the digital elevation model (DEM). Currently available methane inversion products mostly fit broadband structures with polynomials; however, the selection parameters for polynomial order are not well defined, and the broadband structure fitting is not precise enough. In local areas, the high-precision inversion requirements cannot be met by the accuracy of the employed DEM. This paper selects the Qinghai Tibet Plateau region, where the Waliguan Atmospheric Background Reference Observatory is located, as the research area. A higher precision digital elevation model (GLO-30) is used, and a fully connected neural network is used instead of a low order polynomial for "broadband structure" fitting. Furthermore, a "skip" structure is added to the traditional fully connected neural network, and a dropout strategy is used to optimize the network. By Comparing the experimental results with the XCH4 inverted using the global multi resolution terrain elevation data 2010 (GMTED2010) and low order polynomial fitting methods. The results show that the improved fully connected neural network can better fit the broadband spectral structure, and combining it with a higher precision DEM can improve the inversion accuracy of XCH4, with the highest correlation coefficient increased to 0.92. The joint optimization method used can be used for remote sensing inversion of XCH4 in oil and gas production areas, thereby better serving the investigation of methane abnormal emissions in oil and gas production areas.

Key words: WFM-DOAS, XCH4, DEM, broad-band spectra fitting, artificial neural network, satellite remote sensing

中图分类号: 

  • P407.4

图1

反演技术流程图"

图2

单个神经元的激活过程"

图3

SKIP-FCNN结构图"

表1

主要输入数据及来源"

数据 来源
短波红外光谱数据 S-5P
太阳光谱 S-5P
观测几何(观测时间、太阳天顶角、仪器天顶角、方位角、高度、经纬度等) S-5P
温度, 气压, CH4、H2O、CO柱浓度先验廓线数据、干空气柱浓度 美国标准廓线/ ECMWF
吸收气体与干扰气体、标准吸收截面 HITRAN数据库
仪器函数 S-5P
DEM GMTED2010 / GLO-30
AOD、albedo MODIS

图4

瓦里关站点区域两种DEM分布图"

图5

瓦里关站点区域GMTED2010与GLO-30高程差分布图"

图6

基于多项式与SKIP-FCNN的光谱拟合结果对比"

图7

逐天反演结果与瓦里关站点数据对比"

图8

逐月平均反演结果与瓦里关站点数据对比"

1 MILLER S M , MICHALAK A M , DETMERS R G , et al. China's coal mine methane regulations have not curbed growing emissions[J]. Nature Communication, 2019, 10 (1): 303.
doi: 10.1038/s41467-018-07891-7
2 姚璐, 杨东旭, 蔡兆男, 等. 面向我国碳中和、碳达峰的大气甲烷观测卫星现状与发展趋势分析[J]. 大气科学, 2022, 46 (6): 1469- 1483.
YAO Lu , YANG Dongxu , CAI Zhaonan , et al. Status and Trend Analysis of Atmospheric Methane Satellite Measurement for Carbon Neutrality and Carbon Peaking in China[J]. Chinese Journal of Atmospheric Sciences, 2022, 46 (6): 1469- 1483.
3 杨晓钰, 王中挺, 潘光, 等. 卫星遥感温室气体的大气观测技术进展[J]. 大气与环境光学学报, 2022, 17 (6): 581- 597.
doi: 10.3969/j.issn.1673-6141.2022.06.001
YANG Xiaoyu , WANG Zhongting , PAN Guang , et al. Advances in atmospheric observation techniques for greenhouse gases by satellite remote sensing[J]. Journal of Atmospheric and Environmental Optics, 2022, 17 (6): 581- 597.
doi: 10.3969/j.issn.1673-6141.2022.06.001
4 DANIEL J V , DANIEL JJ , DYLAN J , et al. Quantifying time-averaged methane emissions from individual coal mine vents with GHGSat-D satellite observations[J]. Environmental Science & Technology, 2020, 54 (16): 10246- 10253.
5 刘良云, 陈良富, 刘毅, 等. 全球碳盘点卫星遥感监测方法、进展与挑战[J]. 遥感学报, 2022, 26 (2): 243- 267.
LIU Liangyun , CHEN Liangfu , LIU Yi , et al. Satellite remote sensing for global stocktaking: methods, progress and perspectives[J]. National Remote Sensing Bulletin, 2022, 26 (2): 243- 267.
6 刘毅, 王婧, 车轲, 等. 温室气体的卫星遥感——进展与趋势[J]. 遥感学报, 2021, 25 (1): 53- 64.
LIU Yi , WANG Jing , CHE Ke , et al. Satellite remote sensing of greenhouse gases: progress and trends[J]. National Remote Sensing Bulletin, 2021, 25 (1): 53- 64.
7 JONATHAN L, NICO R, ROB V S, et al. Input/output data specification for the TROPOMI L01b data processor[Z]. The Royal Netherlands Meteorological Institute KNMI. 2022.
8 MICHAEL B , VLADIMIR V R , JOHN P B . A near-infrared optimized DOAS method for thefast global retrieval of atmospheric CH4, CO, CO2, H2O, and N2O total column amounts from SCIAMACHY Envisat-1 nadir radiances[J]. Journal of Geophysical Research, 2000, 15, 15231- 15245.
9 OLIVER S , MICHAEL B , MAXIMILIAN R , et al. A scientific algorithm to simultaneously retrieve carbon monoxide and methane from TROPOMI onboard Sentinel-5 Precursor[J]. Atmospheric Measurement Techniques, 2019, 12 (12): 6771- 6802.
10 孙友文, 谢品华, 徐晋, 等. 采用加权函数修正的差分光学吸收光谱反演环境大气中的CO2垂直柱浓度[J]. 物理学报, 2013, 62 (13): 1- 10.
SUN Youwen , XIE Pinhua , XU Jin , et al. Measurement of atmospheric CO2 vertical column density using weighting function modified differential optical absorption spectroscopy[J]. Acta Phys Sin, 2013, 62 (13): 1- 10.
11 OLIVER S. Algorithmtheoretical basis document (ATBD), TROPOMI WFM-DOAS (TROPOMI/WFMD) XCH4[Z], Institute of Environmental Physics, University of Bremen, 2022.
12 OLIVER S , MICHAEL B , JONAS H , et al. Advances in retrieving XCH4 and XCO from Sentinel-5 precursor: improvements in the scientific TROPOMI/WFMD algorithm[J]. Atmospheric Measurement Techniques, 2023, 16, 669- 694.
13 BEREND J S , JOANNES D M , PIETER B , et al. Automated detection and monitoring of methane super-emitters using satellite data[J]. Atmospheric Chemistry and Physics, 2023, 16, 1- 47.
14 JONAS H , OLIVER S , MICHAEL B , et al. On the influence of underlying elevation data on Sentinel-5precursor TROPOMI satellite methane retrievals over Greenland[J]. Atmospheric Measurement Techniques, 2022, 15, 4063- 4074.
15 王汝雯. 基于WFM-DOAS温室气体柱浓度反演方法及应用研究[D]. 中国科学技术大学, 2019.
WANG Ruwen. WFM-DOAS algorithm research and application on greenhouse gases column density retrieval[D]. University of Science and Technology of China, 2019.
16 Hasekamp O, Lorente A, Haili H, et al. Algorithm theoretical baseline document for Sentinel-5 precursor methane retrieval[Z]. Netherlands Institute for Space Research, SRON, 2019.
17 杨倩, 官莉, 陶法, 等. 中国5个大气本底站观测的CH4浓度变化规律[J]. 环境科学与技术, 2018, 41 (6): 1- 7.
YANG Qian , GUAN Li , TAO Fa , et al. Changes of CH4 concentrations obtained by ground-based observations at five atmospheric background stations in China[J]. Environmental Science & Technology, 2018, 41 (6): 1- 7.
18 LIANG Miao. Atmospheric CH4 at Mt. Waliguan by China meteorological administration[DB/OL]. (2023-02-01)[2024-04-08]. https://gaw.kishou.go.jp/.
19 闻言, 徐俊, 高伟, 等. 基于全连接神经网络的旋挖钻机缓冲平衡阀故障诊断[J]. 液压与气动, 2022, 46 (12): 152- 159.
WEN Yan , XU Jun , GAO Wei , et al. Fault diagnosis of cushion balance valve of rotary drilling rig based on fully connected neural network[J]. Chinese Hydraulics & Pneumatics, 2022, 46 (12): 152- 159.
20 LINLU M , VLADIMIR R , ALEXEI R , et al. SCIATRAN software package(V4.6): update and further development of aerosol, clouds, surface reflectance databases and models[J]. Geoscientific Model Development, 2022, 16, 1511- 1536.
21 DANIELSONJ J, GESCHD B. Global multi-resolution terrain elevation data 2010 (GMTED2010)[DB/OL]. U.S. Geological Survey, 2011[2024-04-08]. https://doi.org/10.3133/ofr20111073.
22 European Space Agency, Sinergise. Copernicus global digital elevation model, distributed by open topography[DB/OL]. 2021[2024-04-08]. https://doi.org/10.5069/G9028PQB.
[1] 温晓,刘琪,高振,曾维新,吕咸青. 局部非侵入式约化基模型在瑞利-泰勒不稳定中的应用[J]. 《山东大学学报(理学版)》, 2020, 55(2): 109-117.
[2] 宋元章,李洪雨,陈媛,王俊杰. 基于分形与自适应数据融合的P2P botnet检测方法[J]. 山东大学学报（理学版）, 2017, 52(3): 74-81.
[3] 堵锡华,史小琴,冯长君,李亮. 基于野韭菜挥发性成分的色谱保留指数神经网络预测[J]. 山东大学学报（理学版）, 2014, 49(1): 50-53.
[4] 柳扬. 布尔函数代数免疫阶的计算[J]. J4, 2010, 45(3): 55-60.
[5] 贺晓丽1, 伏文清2. L-预拓扑空间的局部连通性:可乘性与L-好的推广[J]. J4, 2010, 45(10): 78-82.
[6] 贺晓丽,伏文清,李生刚 . L-预拓扑空间的局部连通性[J]. J4, 2007, 42(8): 58-61 .
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
[1] 王 兵 . 拟无爪图的性质[J]. J4, 2007, 42(10): 111 -113 .
[2] 朱焱,侯建锋,王纪辉 . 图的粘合运算与韧度和孤立韧度的关系[J]. J4, 2006, 41(5): 59 -62 .
[3] 江雪莲,石洪波*. 产生式与判别式组合分类器学习算法[J]. J4, 2010, 45(7): 7 -12 .
[4] 彭艳芬,李宝宗,刘天宝 . 有机气体麻醉活性的构效关系研究[J]. J4, 2006, 41(5): 148 -150 .
[5] 孙守斌,孟广武,赵 峰 . 序同态的Dα-连续性[J]. J4, 2007, 42(7): 49 -53 .
[6] 郭亭,鲍晓明 . P137G点突变对嗜热细菌木糖异构酶酶活性及热稳定性的影响[J]. J4, 2006, 41(6): 145 -148 .
[7] 戴珍香,李曙光,亓兴勤 . 波分复用星形单跳网中3信道的传输调度问题[J]. J4, 2007, 42(2): 46 -50 .
[8] 刘海英 马成刚 王志平. 刺图乘积上的Graham猜想[J]. J4, 2009, 44(8): 25 -30 .
[9] 杨伦，徐正刚，王慧*，陈其美，陈伟，胡艳霞，石元，祝洪磊，曾勇庆*. RNA干扰沉默PID1基因在C2C12细胞中表达的研究[J]. J4, 2013, 48(1): 36 -42 .
[10] 刘艳萍,吴群英. 优化权重下高斯序列最大值几乎处处中心极限定理[J]. 山东大学学报（理学版）, 2014, 49(05): 50 -53 .
版权所有 © 2018 《山东大学学报(理学版)》编辑部　
地址: 山东省济南市山大南路27号山东大学中心校区（250100）　电话: +86-0531-88366917　E-mail: xblxb@sdu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发