Please wait a minute...
img

官方微信

遥感技术与应用  2020, Vol. 35 Issue (3): 615-622    DOI: 10.11873/j.issn.1004-0323.2020.3.0615
模型与反演     
基于3种机器学习法的太阳辐射模拟研究
李净(),温松楠()
西北师范大学 地理与环境科学学院,甘肃 兰州 730070
Simulation of Solar Radiation based on Three Machine Learning Methods
Jing Li(),Songnan Wen()
The College of Geographical and Environmental Science, Northwest Normal University, Lanzhou 730070, China
 全文: PDF(3463 KB)   HTML
摘要:

定量模拟太阳辐射对认识黄土高原区气候变化至关重要,现有研究表明机器学习可以很好地模拟太阳辐射,但不同的机器学习法在不同区域模拟精度不同,为了实现黄土高原区太阳辐射数据的最优模拟,从而为农作物模型、水文模型以及气候变化模型提供精度更高的太阳辐射数据。基于随机森林(RF,Random Forest)、人工神经网络(ANN,Artificial Neural Network)和支持向量机(SVM,Support Vector Machine)3种机器学习法来模拟黄土高原地区的太阳辐射并对这3种算法进行比较研究,选取了2003~2009年14个辐射站点和2010~2016年10个辐射站点的实测数据和对应参数气压、云量、云光学厚度、臭氧、可降水水汽以及DEM、坡度、坡向作为模型的训练数据,随机选取2010~2016年4个辐射站点的太阳辐射实测数据对模拟结果进行验证。验证结果表明:RF模型在黄土高原及周边地区的模拟效果最优,平均偏差(MBE)为-0.17 MJ·m-2,均方根误差(RMSE)为1.48 MJ·m-2,拟合优度达到0.96。研究结果表明:RF模型与气象数据及遥感数据结合能够有效解决黄土高原无辐射观测区的太阳辐射模拟问题,对区域太阳辐射的研究具有重要意义。

关键词: 太阳辐射随机森林(RF)人工神经网络(ANN)支持向量机(SVM)遥感    
Abstract:

Quantitative simulation of solar radiation is essential for understanding climate change 0n the Loess Plateau, Many machine learning methods were developed to estimate solar radiation well, but different machine learning methods have different simulation accuracy in different regions, In order to achieve optimal simulation of solar radiation on the Loess Plateau, this provides more higher precision solar radiation data for crop models, hydrological models, and climate change models. In this study, three machine learning methods, including Random Forest (RF), Artificial Neural Network (ANN) and Support Vector Machine (SVM), were applied to estimate solar radiation on the Loess Plateau, three machine learning methods were trained using ground measurements at fourteen radiation sites from 2003 to 2009 and ten radiation sites from 2010 to 2016 and corresponding parameter pressure, cloud fraction, cloud optical thickness, ozone, precipitation water vapor and DEM, slope, and aspect to train the three model, The solar radiation estimates based on three machine learning methods were evaluated using ground measurements at four radiation sites from 2010 to 2016. The validation results show that the RF model has the best simulation effect on the Loess Plateau and surrounding areas. The average deviation is -0.17 MJ·m-2, the root mean square error is 1.48 MJ·m-2, and has a good fit of 0.96. The results show that combined RF model and meteorological data and remote sensing data can effectively solve the problem about solar radiation simulation on the non-radiation observation area of the Loess Plateau, which is of great significance to the research of regional solar radiation.

Key words: Solar radiation    Random Forest(RF)    Artificial Neural Network(ANN)    Support Vector Machine(SVM)    Remote sensing
收稿日期: 2019-06-04 出版日期: 2020-07-10
ZTFLH:  P422.1  
基金资助: 国家自然科学基金项目(41261016);西北师范大学青年教师科研能力提升计划项目(NWNU-LKQN-14-4)
通讯作者: 温松楠     E-mail: li_jinger@163.com;18893111471@163.com
作者简介: 李 净(1978-),女,甘肃会宁人,副教授,硕士生导师,主要从事定量遥感研究。E-mail:li_jinger@163.com
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
李净
温松楠

引用本文:

李净,温松楠. 基于3种机器学习法的太阳辐射模拟研究[J]. 遥感技术与应用, 2020, 35(3): 615-622.

Jing Li,Songnan Wen. Simulation of Solar Radiation based on Three Machine Learning Methods. Remote Sensing Technology and Application, 2020, 35(3): 615-622.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2020.3.0615        http://www.rsta.ac.cn/CN/Y2020/V35/I3/615

图1  黄土高原地区辐射站点图
训练误差验证误差
模型参数MBE/MJ·m-2RMSE/MJ·m-2RMBE/MJ·m-2RMSE/MJ·m-2R
RF最大特征数8、最大深度100.252.430.92-0.171.480.96
SVM核函数为RBF、gamma为0.02、C为100000.162.660.841.552.630.92
ANN训练函数为trainlm、神经网络结构为8-20-1-1.344.270.73-1.144.010.75
表1  模型误差
图2  太阳辐射月均值模拟值与实测值对比
图3  模拟验证误差
1 Sun H, Zhao N, Zeng X, et al. Study of Solar Radiation Prediction and Modeling of Relationships between Solar Radiation and Meteorological Variables[J]. Energy Conversion and Management, 2015, 105: 880-890. doi:10.1016/j.enconman.2015.08.045.
doi: 10.1016/j.enconman.2015.08.045
2 Chen J L, Li G S, Wu S J. Assessing the Potential of Support Vector Machine for Estimating Daily Solar Radiation Using Sunshine Duration[J]. Energy Convers Manage, 2013,75:311-318.
3 Shi Guoping, Qiu Xinfa, Zeng Yan. Distributed Simulation of Three Solar Radiation Starting Data in China[J]. Geographical Science, 2013, 33(4): 385-392.
3 施国萍, 邱新法, 曾燕. 中国三种太阳辐射起始数据分布式模拟[J]. 地理科学, 2013, 33(4): 385-392.
4 Li Weiwei, Letu Husi, Chen Hongbin. Calculation of Surface Solar Radiation Under Different Cloud Conditions Using MODIS Data[J]. Remote Sensing Technology and Application, 2017,32(4): 643-650.
4 黎微微, 胡斯勒图, 陈洪滨. 利用 MODIS 资料计算不同云天条件下的地表太阳辐射[J]. 遥感技术与应用, 2017,32(4): 643-650.
5 Journée M, Bertrand C. Improving the Spatio-temporal Distribution of Surface Solar Radiation Data by Merging Ground and Satellite Measurements[J]. Remote Sensing of Environment, 2010, 114(11): 2692-2704.
6 Zhang T, Stackhouse Jr P W, Gupta S K, et al. The Validation of the GEWEX SRB Surface Shortwave Flux Data Products Using BSRN Measurements: A Systematic Quality Control, Production and Application Approach[J]. Journal of Quantitative Spectroscopy and Radiative Transfer, 2013, 122: 127-140.
7 Huang G, Liang S, Lu N, et al. Toward a Broadband Parameterization Scheme for Estimating Surface Solar Irradiance: Development and Preliminary Results on MODIS Products[J]. Journal of Geophysical Research: Atmospheres, 2018, 123(21): 12180-12193.
8 Gueymard C A, Myers D R. Evaluation of Conventional and High-performance Routine Solar Radiation Measurements for Improved Solar Resource, Climatological Trends, and Radiative Modeling[J]. Solar Energy, 2009, 83(2): 171-185.
9 Angstrom A. Solar and Terrestrial Radiation. Report to the International Commission for Solar Research on Actinometric Investigations of Solar and Atmospheric Radiation[J]. Quarterly Journal of the Royal Meteorological Society, 1924, 50(210): 121-126.
10 Li R, Zhao L, Wu T, et al. Temporal and Spatial Variations of Global Solar Radiation over the Qinghai-Tibetan Plateau during the Past 40 Years[J]. Theoretical and Applied Climatology, 2013, 113(3-4): 573-583.
11 Mummadisetty B C, Puri A, Sharifahmadian E, et al. A hybrid Method for Compression of Solar Radiation Data Using Neural Networks[J]. International Journal of Communications, Network and System Sciences, 2015, 8(6): 217. doi:10.4236/ijcns.2015.86022.
doi: 10.4236/ijcns.2015.86022
12 Belaid S, Mellit A. Prediction of Daily and Mean Monthly Global Solar Radiation Using Support Vector Machine in an Arid Climate[J]. Energy Conversion and Management, 2016, 118: 105-118.
13 Chen R, Ersi K, Yang J, et al. Validation of Five Global Radiation Models with Measured Daily Data in China[J]. Energy Conversion and Management, 2004, 45(11-12): 1759-1769.
14 Peng Z, Letu H, Wang T, et al. Estimation of Shortwave Solar Radiation Using the Artificial Neural Network from Himawari-8 Satellite Imagery over China[J]. Journal of Quantitative Spectroscopy and Radiative Transfer, 2020, 240: 106672.
15 Zhu X, Qiu X, Zeng Y, et al. A Remote Sensing Model to Estimate Sunshine Duration in the Ningxia Hui Autonomous Region, China[J]. Journal of Meteorological Research, 2015, 29(1): 144-154.
16 Sun H, Gui D, Yan B, et al. Assessing the Potential of Random Forest Method for Estimating Solar Radiation Using Air Pollution Index[J]. Energy Conversion and Management, 2016, 119: 121-129.
17 Benali L, Notton G, Fouilloy A, et al. Solar Radiation Forecasting Using Artificial Neural Network and Random Forest Methods: Application to Normal Beam, Horizontal Diffuse and Global Components[J]. Renewable Energy, 2019, 132: 871-884.
18 Liang Yitong, Liu Kequn, Xia Zhihong. Estimation of Solar Radiation Using FY-2C Satellite Data[J]. Meteorological Science and Technology, 2009, 37(2): 234-238.
18 梁益同, 刘可群, 夏智宏. 利用FY-2C卫星资料估算太阳辐射研究[J]. 气象科技, 2009, 37(2):234-238.
19 Deo R C, Şahin M. Forecasting Long-term Global Solar Radiation with an ANN Algorithm Coupled with Satellite-derived (MODIS) Land Surface Temperature (LST) for Regional Locations in Queensland[J]. Renewable and Sustainable Energy Reviews, 2017, 72: 828-848.
20 Fallahi S, Amanollahi J, Tzanis C G, et al. Estimating Solar Radiation Using NOAA/AVHRR and Ground Measurement Data[J]. Atmospheric Research, 2018, 199: 93-102.
21 Yao W, Zhang C, Hao H, et al. A Support Vector Machine Approach to Estimate Global Solar Radiation with the Influence of fog and Haze[J]. Renewable Energy, 2018, 128: 155-162.
22 Li Jing, Wang Dan, Feng Wei. Simulation of Solar Radiation based on MODIS Remote Sensing Products and Neural Networks[J]. Geographical Science, 2017, 37(6): 912-919.
22 李净, 王丹, 冯姣姣. 基于 MODIS 遥感产品和神经网络模拟太阳辐射[J]. 地理科学, 2017, 37(6): 912-919.
23 Breiman L. Random Forests[J]. Machine Learning, 2001, 45(1): 5-32.
24 Breiman L. Bagging Preditors[J]. Machine Learning, 1996, 24(2):123-140.
25 Fang Kuangnan, Wu Jianbin, Zhu Jianping, et al. Review of Random Forest Methods Research[J]. Journal of Statistics and Information, 2011, 26(3): 32-38.
25 方匡南, 吴见彬, 朱建平, 等. 随机森林方法研究综述[J]. 统计与信息论坛, 2011, 26(3):32-38.
26 Zhao Lingling, Weng Suming, Zeng Huajun, et al. Nuclear Method for Pattern Analysis [M]. Beijing: Mechanical Engineering Press, 2006.赵玲玲, 翁苏明, 曾华军, 等. 模式分析的核方法[M]. 北京: 机械工程出版社,2006.
27 Gao Wei. Principles and Simulation Examples of Artificial Neural Networks [M]. Beijing: Mechanical Engineering Press, 2003.
27 高隽.人工神经网络原理及仿真实例[M].北京: 机械工程出版社,2003.
28 Vapnik V N. The Nature of Statistical Learning Theory [M].NY: Springer-Verlag, 1995.
29 Cortes C,Vapnik V. Support-vector Network [J]. Machine Learning,1995,20(3):273-297.
30 Quej V H, Almorox J, Ibrakhimov M, et al. Empirical Models for Estimating Daily Global Solar Radiation in Yucatán Peninsula, Mexico[J]. Energy Conversion and Management, 2016, 110: 448-456.
31 Meenal R, Selvakumar A I. Assessment of SVM, Empirical and ANN based Solar Radiation Prediction Models with Most Influencing Input Parameters[J]. Renewable Energy, 2018, 121: 324-343.
32 Feng Jiaojiao, Wang Weizhen, Li Jing, et al. Analysis of Solar Radiation Simulation and Time-space Change in East China based on BP Neural Network[J]. Remote Sensing Technology and Application, 2018, 33(5): 881-889.
32 冯姣姣, 王维真, 李净, 等. 基于BP神经网络的华东地区太阳辐射模拟及时空变化分析[J]. 遥感技术与应用, 2018, 33(5):881-889.
33 Wei Y, Zhang X, Hou N, et al. Estimation of Surface Downward Shortwave Radiation over China from AVHRR Data based on Four Machine Learning Methods[J]. Solar Energy, 2019, 177: 32-46.
[1] 李智礼,匡文慧,张澍. 近70 a天津主城区城市土地利用/覆盖变化遥感监测与时空分析[J]. 遥感技术与应用, 2020, 35(3): 527-536.
[2] 史姝姝,匡文慧,董斯齐. 21世纪以来西安城乡梯度土地覆盖变化及对城市热岛影响时空特征[J]. 遥感技术与应用, 2020, 35(3): 537-547.
[3] 虞瑶,苏红军,姚文静. 基于Boosting的高光谱遥感切空间协同表示集成学习方法[J]. 遥感技术与应用, 2020, 35(3): 634-644.
[4] 马彩虹,关琳琳,陈甫,王大成,刘建波. 基于内容的遥感图像变化信息检索概念模型设计[J]. 遥感技术与应用, 2020, 35(3): 685-693.
[5] 岳昂,曾庆伟,王怀警. 于桥水库蓝藻水华遥感长时序监测研究[J]. 遥感技术与应用, 2020, 35(3): 694-701.
[6] 谷祥辉,张英,桑会勇,翟亮,李少军. 基于哨兵2时间序列组合植被指数的作物分类研究[J]. 遥感技术与应用, 2020, 35(3): 702-711.
[7] 寇勇,王宁练,陈安安,刘凯. 1993~2016年喜马拉雅山西段杰纳布流域冰川变化遥感监测[J]. 遥感技术与应用, 2020, 35(3): 712-722.
[8] 张平,孙强强,张亚萍,孙丹峰,刘顺喜. 基于宽波段遥感光谱混合分解的干旱区土地退化监测评价理论和技术方法探讨[J]. 遥感技术与应用, 2020, 35(2): 497-508.
[9] 程娟,肖青,闻建光,唐勇,游冬琴,卞尊健,吴胜标,郝大磊,钟守熠. 地物波谱数据库应用方法及遥感应用现状[J]. 遥感技术与应用, 2020, 35(2): 267-286.
[10] 谭美宝,冉有华,苏阳,李新,杜得彦,廉耀康. 黑河流域2001~2017年植被变化特征及其可延续性评价[J]. 遥感技术与应用, 2020, 35(2): 335-344.
[11] 纪甫江,蒙继华,方慧婷. 基于CASA-WOFOST耦合模型的大豆单产遥感估算研究[J]. 遥感技术与应用, 2020, 35(2): 406-415.
[12] 唐廷元,付波霖,何素云,娄佩卿,闭璐. 基于GF-1和Sentinel-1A的漓江流域典型地物信息提取[J]. 遥感技术与应用, 2020, 35(2): 448-457.
[13] 李宜展, 李泽霞, 刘细文, 魏韧, 郭世杰, 董璐. 遥感科学数据应用态势分析[J]. 遥感技术与应用, 2020, 35(1): 1-12.
[14] 张新平,乔治,李皓,闫杰,张芳芳,赵栋锋,王得祥,康海斌,杨航,冯扬. 基于Landsat影像和不规则梯形方法遥感反演延安城市森林表层土壤水分[J]. 遥感技术与应用, 2020, 35(1): 120-131.
[15] 李雷,郑兴明,赵凯,李晓峰,王广蕊. 基于CCI土壤水分产品的干旱指数精度评价及其对东北地区粮食产量的影响[J]. 遥感技术与应用, 2020, 35(1): 111-119.