首页 > 论文 > 光学学报 > 38卷 > 10期(pp:1030001--1)

基于竞争适应重加权采样算法耦合机器学习的土壤含水量估算

Estimation of Soil Moisture Content Based on Competitive Adaptive Reweighted Sampling Algorithm Coupled with Machine Learning

  • 摘要
  • 论文信息
  • 参考文献
  • 被引情况
  • PDF全文
分享:

摘要

土壤含水量是干旱区地表水-热-溶质耦合运移的关键指标;以干旱区典型样点实测土壤含水量及其室内可见光-近红外光谱数据作为数据集,通过蒙特卡罗交叉验证确定77个有效样本;基于竞争适应重加权采样算法筛选出最优光谱变量子集,利用3种机器学习方法——BP神经网络、随机森林回归和极限学习机建立土壤含水量预测模型,进而实现土壤含水量估算模型的优选。结果表明:竞争适应重加权采样算法能有效剔除无关变量,从2151个光谱波段中优选出20个特征波段,其中R1848与土壤含水量的最大相关系数为0.531;引入偏最小二乘模型和机器学习方法进行对比,分析发现机器学习方法的预测结果比偏最小二乘模型更高;分析比较BP神经网络、随机森林回归和极限学习机的建模结果可知:极限学习机模型建模在机器学习方法中的效果最佳,决定系数R2=0.918,均方根误差RMSE=0.015,相对分析误差RPD=3.123,四分位数间隔RPIQ=3.325;机器学习能显著提升光谱建模反演土壤含水量的精度和稳定性,显示出其在非线性问题中具有很强的透析力和较好的模型稳健性,针对干旱区土壤水分的精准预测和定量估算具有可行性,可为干旱区土壤墒情、精准农业等研究提供科学参考。

Abstract

Soil moisture content is an important indicator that reflects the coupled surface water-heat-solute transport in arid regions. The visible and near-infrared spectroscopy has been widely used for soil moisture content prediction owing to its rapid response. The soil moisture content and corresponding spectral data are obtained in the laboratory; then, the calibration datasets (n=77) are selected using Monte Carlo cross-validation algorithm. The competitive adaptive reweighted sampling algorithm is used to optimize spectral variables. Three machine learning algorithms, namely back propagation neural network, random forest regression, and extreme learning machine are used to construct predicting models. The results reveal that competitive adaptive reweighted sampling algorithm can effectively filter and eliminate massive irrelevant variables. Herein, a total of 20 feature bands are divided from all spectral bands, where the band of R1848 is the most prominent (the maximum correlation coefficient is 0.531). The performance of models based on machine learning algorithms is superior to those based on partial least squares regression, with the optimal prediction of the coefficient of determination (R2), root mean square error of prediction (RMSE), residual predictive deviation (RPD), and ratio of performance to interquartile range (RPIQ). Compared with the predictive effects of all the models, the extreme learning machine-based predicting model is the most effective (R2=0.918, RMSE=0.015, RPD=3.123, and RPIQ=3.325). Compared with common linear models, the machine learning algorithms can effectively improve the precision and stability of the quantitative estimation of soil moisture content. The results provide scientific guidance and baseline data for the accurate monitoring of soil moisture content and precision agriculture in arid regions.

Newport宣传-MKS新实验室计划
补充资料

中图分类号:O436

DOI:10.3788/aos201838.1030001

所属栏目:光谱学

基金项目:自治区重点实验室专项基金(2016D03001)、国家自然科学基金(41771470,U1303381,41661046)

收稿日期:2018-04-28

修改稿日期:2018-05-07

网络出版日期:2018-05-12

作者单位    点击查看

葛翔宇:新疆大学资源与环境科学学院, 新疆 乌鲁木齐 830046新疆大学绿洲生态教育部重点实验室, 新疆 乌鲁木齐 830046新疆大学智慧城市与环境建模自治区普通高校重点实验室, 新疆 乌鲁木齐 830046
丁建丽:新疆大学资源与环境科学学院, 新疆 乌鲁木齐 830046新疆大学绿洲生态教育部重点实验室, 新疆 乌鲁木齐 830046新疆大学智慧城市与环境建模自治区普通高校重点实验室, 新疆 乌鲁木齐 830046
王敬哲:新疆大学资源与环境科学学院, 新疆 乌鲁木齐 830046新疆大学绿洲生态教育部重点实验室, 新疆 乌鲁木齐 830046新疆大学智慧城市与环境建模自治区普通高校重点实验室, 新疆 乌鲁木齐 830046
王飞:新疆大学资源与环境科学学院, 新疆 乌鲁木齐 830046新疆大学绿洲生态教育部重点实验室, 新疆 乌鲁木齐 830046新疆大学智慧城市与环境建模自治区普通高校重点实验室, 新疆 乌鲁木齐 830046
蔡亮红:新疆大学资源与环境科学学院, 新疆 乌鲁木齐 830046新疆大学绿洲生态教育部重点实验室, 新疆 乌鲁木齐 830046新疆大学智慧城市与环境建模自治区普通高校重点实验室, 新疆 乌鲁木齐 830046
孙慧兰:新疆师范大学地理科学与旅游学院, 新疆 乌鲁木齐 830054

联系人作者:丁建丽(watarid@xju.edu.cn); 葛翔宇(xiangyu_gexj@163.com);

【1】Kumar S V, Dirmeyer P A, Peters-Lidard C D, et al. Information theoretic evaluation of satellite soil moisture retrievals[J]. Remote Sensing of Environment, 2018, 204: 392-400.

【2】Cai L H, Ding J L. Inversion of soil moisture content based on hyperspectral multi-scale decomposition[J]. Laser & Optoelectronics Progress, 2018, 55(1): 013001.
蔡亮红, 丁建丽. 基于高光谱多尺度分解的土壤含水量反演[J]. 激光与光电子学进展, 2018, 55(1): 013001.

【3】Yu L, Zhu Y X, Hong Y S, et al. Determination of soil moisture content by hyperspectral technology with CARS algorithm[J]. Transactions of the Chinese Society of Agricultural Engineering, 2016, 32(22): 138-145.
于雷, 朱亚星, 洪永胜, 等. 高光谱技术结合CARS算法预测土壤水分含量[J]. 农业工程学报, 2016, 32(22): 138-145.

【4】Xu C, Zeng W Z, Huang J S, et al. Prediction of soil moisture content and soil salt concentration from hyperspectral laboratory and field data[J]. Remote Sensing, 2016, 8(1): 42.

【5】Oltra-Carrió R, Baup F, Fabre S, et al. Improvement of soil moisture retrieval from hyperspectral VNIR-SWIR data using clay content information: from laboratory to field experiments[J]. Remote Sensing, 2015, 7(3): 3184-3205.

【6】Zhu Y X, Yu L, Hong Y S, et al. Hyperspectral features and wavelength variables selection methods of soil organic matter[J]. Scientia Agricultura Sinica, 2017, 50(22): 4325-4337.
朱亚星, 于雷, 洪永胜, 等. 土壤有机质高光谱特征与波长变量优选方法[J]. 中国农业科学, 2017, 50(22): 4325-4337.

【7】Kawamura K, Tsujimoto Y, Rabenarivo M, et al. Vis-NIR spectroscopy and PLS regression with waveband selection for estimating the total C and N of paddy soils in Madagascar[J]. Remote Sensing, 2017, 9(10): 1081.

【8】Zhang X L, Zhang F, Zhang H W, et al. Optimization of soil salt inversion model based on spectral transformation from hyperspectral index[J]. Transactions of the Chinese Society of Agricultural Engineering, 2018, 34(1): 110-117.
张贤龙, 张飞, 张海威, 等. 基于光谱变换的高光谱指数土壤盐分反演模型优选[J]. 农业工程学报, 2018, 34(1): 110-117.

【9】Diao W Y, Liu G, Hu K L. Estimation of soil water content based on hyperspectral features and the ANN model[J]. Spectroscopy and Spectral Analysis, 2017, 37(3): 841-846.
刁万英, 刘刚, 胡克林. 基于高光谱特征与人工神经网络模型对土壤含水量估算[J]. 光谱学与光谱分析, 2017, 37(3): 841-846.

【10】Belgiu M, Drgu瘙塅 L. Random forest in remote sensing: a review of applications and future directions[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2016, 114: 24-31.

【11】Khosravi V, Ardejani F D, Yousefi S, et al. Monitoring soil lead and zinc contents via combination of spectroscopy with extreme learning machine and other data mining methods[J]. Geoderma, 2018, 318: 29-41.

【12】Nawar S, Mouazen A M. Comparison between random forests, artificial neural networks and gradient boosted machines methods of on-line vis-NIR spectroscopy measurements of soil total nitrogen and total carbon[J]. Sensors, 2017, 17(10): 2428.

【13】Liang L, Di L P, Zhang L P, et al. Estimation of crop LAI using hyperspectral vegetation indices and a hybrid inversion method[J]. Remote Sensing of Environment, 2015, 165: 123-134.

【14】He B Z, Ding J L, Wang F, et al. Research on data mining of salinization information based on phenological characters[J]. Acta Ecologica Sinica, 2017, 37(9): 3133-3148.
何宝忠, 丁建丽, 王飞, 等. 基于物候特征的盐渍化信息数据挖掘研究[J]. 生态学报, 2017, 37(9): 3133-3148.

【15】Vohland M, Ludwig M, Thiele-Bruhn S, et al. Determination of soil properties with visible to near- and mid-infrared spectroscopy: effects of spectral variable selection[J]. Geoderma, 2014, 223/224/225: 88-96.

【16】Tan K, Wang H M, Zhang Q Q, et al. An improved estimation model for soil heavy metal(loid) concentration retrieval in mining areas using reflectance spectroscopy[J]. Journal of Soils and Sediments, 2018, 18: 2008-2022.

【17】Tang G, Huang Y, Tian K D, et al. A new spectral variable selection pattern using competitive adaptive reweighted sampling combined with successive projections algorithm[J]. Analyst, 2014, 139(19): 4894-4902.

【18】Wijewardane N K, Ge Y F, Morgan C L S. Moisture insensitive prediction of soil properties from VNIR reflectance spectra based on external parameter orthogonalization[J]. Geoderma, 2016, 267: 92-101.

【19】Liu Y Q, Chen H Y, Wang R Y, et al. Quantitative analysis of soil salt and its main ions based on visible/near infrared spectroscopy in estuary area of Yellow River[J]. Scientia Agricultura Sinica, 2016, 49(10): 1925-1935.
刘亚秋, 陈红艳, 王瑞燕, 等. 基于可见/近红外光谱的黄河口区土壤盐分及其主要离子的定量分析[J]. 中国农业科学, 2016, 49(10): 1925-1935.

【20】Huang G, Huang G B, Song S J, et al. Trends in extreme learning machines: a review[J]. Neural Networks, 2015, 61: 32-48.

【21】Cheng S X, Kong W W, Zhang C, et al. Variety recognition of Chinese cabbage seeds by hyperspectral imaging combined with machine learning[J]. Spectroscopy and Spectral Analysis, 2014, 34(9): 2519-2522.
程术希, 孔汶汶, 张初, 等. 高光谱与机器学习相结合的大白菜种子品种鉴别研究[J]. 光谱学与光谱分析, 2014, 34(9): 2519-2522.

【22】Wang J Z, Ding J L, Abulimiti A, et al. Quantitative estimation of soil salinity by means of different modeling methods and visible-near infrared (VIS-NIR) spectroscopy, Ebinur Lake Wetland, Northwest China[J]. PeerJ, 2018, 6: e4703.

【23】Yu X, Liu Q, Wang Y B, et al. Evaluation of MLSR and PLSR for estimating soil element contents using visible/near-infrared spectroscopy in apple orchards on the Jiaodong Peninsula[J]. Catena, 2016, 137: 340-349.

【24】Li Z, Zhang F, Feng H K, et al. Research on the estimation of salt ions of vegetation leaves based on band combination[J]. Acta Optica Sinica, 2017, 37(11): 1128002.
李哲, 张飞, 冯海宽, 等. 基于波段组合的植被叶片盐离子估算研究[J]. 光学学报, 2017, 37(11): 1128002.

引用该论文

Ge Xiangyu,Ding Jianli,Wang Jingzhe,Wang Fei,Cai Lianghong,Sun Huilan. Estimation of Soil Moisture Content Based on Competitive Adaptive Reweighted Sampling Algorithm Coupled with Machine Learning[J]. Acta Optica Sinica, 2018, 38(10): 1030001

葛翔宇,丁建丽,王敬哲,王飞,蔡亮红,孙慧兰. 基于竞争适应重加权采样算法耦合机器学习的土壤含水量估算[J]. 光学学报, 2018, 38(10): 1030001

您的浏览器不支持PDF插件,请使用最新的(Chrome/Fire Fox等)浏览器.或者您还可以点击此处下载该论文PDF