高光谱成像的土壤剖面水分含量反演及制图

doi:10.3964/j.issn.1000-0593(2019)09-2847-08

摘要
参考文献
相关文章 (15)

全文: PDF (3817 KB)
输出: BibTeX | EndNote (RIS)

摘要：传统土壤水分的获取方法仅可获得离散的土壤水分点位数据，难以获得剖面上精细且连续的水分含量分布图。研究了野外条件下利用近红外高光谱（882～1 709 nm）成像反演剖面土壤水分含量（SMC），并实现精细制图的可行性。研究剖面位于江苏省东台市，我们利用近红外高光谱成像仪对剖面进行了5天原位连续观测，共采集了280个土样用于烘干法测定SMC。原始高光谱图像经数字量化值（DN）校正、黑白校正、拼接、几何校正、剪切和掩膜等一系列预处理后，提取各采样点的平均光谱反射率。提取光谱（Raw）经吸光度[LOG₁₀(1/R)]，Savitzky-Golay平滑（SG）、一阶微分（FD）、二阶微分（SD）、多元散射校正（MSC）和标准正态变量（SNV）转换后，采用偏最小二乘回归（PLSR）和最小二乘支持向量机（LS-SVM）方法建立SMC预测模型，并对比分析不同光谱预处理方法与建模方法组合条件下SMC的预测精度。结果表明，光谱反射率随SMC增加逐渐降低，不同光谱预处理方法的预测精度有所差异，除MSC方法外，同一光谱预处理方法的LS-SVM模型预测精度均高于PLSR模型，并且基于LOG₁₀(1/R)光谱的LS-SVM模型对SMC预测精度最高，其建模集的决定系数（R²_c）和均方根误差（RMSE_c）分别为0.96和0.65%，预测集的决定系数（R²_p）、均方根误差（RMSE_p）和相对分析误差（RPD_p）分别为0.88，1.05%和2.88。利用最优模型进行剖面SMC的高空间分辨率精细制图，通过比较SMC反演图中提取的预测值与实测值关系发现预测精度较高（R²: 0.85～0.95, RMSE: 0.94%～1.02%），且两者在剖面中的变化趋势基本一致，说明SMC反演图不仅能很好地反映出土壤水分在整个剖面中毫米级的含量分布信息，也可反映出同一位置处不同天数间的含量差异。因此，利用近红外高光谱成像结合优化的预测模型，能够实现土壤剖面SMC的定量预测及精细制图，有助于快速、有效监测田间剖面土壤水分状况。

关键词：剖面；土壤含水量；高光谱成像；偏最小二乘回归；最小二乘支持向量机；制图

Abstract：Traditional methods for acquiring soil moisture can only provide discrete point data, which arenot so appropriate for finely and continuously mapping soil moisture distribution in soil profile. In this paper, the feasibility of predicting and mapping ofsoil moisture content (SMC) in soil profile was studied using near-infrared hyperspectral imaging in the spectral range of 882～1 709 nm. Two soil profiles, located in Dongtai City of Jiangsu Province, were continuously observed in situ for 5 days by the near-infrared hyperspectral imaging system. A total of 280 soil samples were obtained for later SMC measurement by oven-drying method. After a series of preproces on the acquired raw hyperspectral images, including digital number(DN) correction, reflectance correction, mosaicking, geometric correction, image clipping and masking, the average spectral reflectance of each sampling point in the corrected hyperspectral images was extracted for further analysis. Then the extracted spectra (Raw) were preprocessed by LOG₁₀(1/R), Savitzky-Golay (SG), first derivative (FD), second derivative (SD), multiplicative scatter correction (MSC) and standard normal variate (SNV), and partial least squares regression (PLSR) and least squares support vector machine (LS-SVM) models were developed and comparedfor a selection of optimum prediction model. Results showed that the soil spectral reflectance gradually decreased with the increase of SMC, and different spectral preprocessing methods had different prediction accuracy. Except for the MSC preprocessing method, the prediction accuracy of the LS-SVM model was higher than the PLSR model with the same spectral preprocessing method. The prediction accuracy of the LS-SVM model with LOG₁₀(1/R) preprocessed spectra was highest with R²_c of 0.96 and RMSE_c of 0.65% for calibration, and R²_p of 0.88, RMSE_p of 1.05% and RPD_p of 2.88 for prediction. The optimum model was then applied to produce high spatial resolution maps of SMC in profiles. The prediction accuracy was high (R²: 0.85～0.95, RMSE: 0.94%～1.02%) by comparing the extracted SMC values from prediction maps with the measured values, and both SMC had the samedistribution tendency in profiles, demonstrating that the SMC prediction mapscould well displaynot only the SMC distribution in profiles in the millimeter scale, but also the changes of SMC at different locations in the profile between different days. Thus, the near-infrared hyperspectral imaging technology combined with optimized prediction model could provide a new approach to quantitatively predict and map high spatial resolution images of SMC in soil profiles in situ, which could help to rapidlyand effectively monitorsoil moisture in profiles in the field.

Key words：Profile; Soil moisture content; Hyperspectral imaging; Partial least squares regression; Least squares support vector machine; Mapping

收稿日期: 2018-07-06 修订日期: 2018-12-09

中图分类号:

TP79

基金资助: 国家自然科学基金项目（41401507，41601214）, 中国科学院战略性先导科技专项（XDB15040300），中国科学院南京土壤研究所“一三五”计划和领域前沿项目（ISSASIP1629）资助

通讯作者: 潘贤章 E-mail: panxz@issas.ac.cn

作者简介: 吴士文，1987年生，中国科学院大学博士研究生 e-mail: 1987wushiwen@163.com

引用本文:

吴士文，王昌昆，刘娅，李燕丽，刘杰，徐爱爱，潘恺，李怡春，张芳芳，潘贤章. 高光谱成像的土壤剖面水分含量反演及制图[J]. 光谱学与光谱分析, 2019, 39(09): 2847-2854.
WU Shi-wen, WANG Chang-kun, LIU Ya, LI Yan-li, LIU Jie, XU Ai-ai, PAN Kai, LI Yi-chun, ZHANG Fang-fang, PAN Xian-zhang. Inversion and Mapping of the Moisture Content in Soil Profiles Based on Hyperspectral Imaging Technology. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2019, 39(09): 2847-2854.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2019)09-2847-08 或 https://www.gpxygpfx.com/CN/Y2019/V39/I09/2847

[1] LIU Ya, PAN Xian-zhang, WANG Chang-kun, et al(刘娅，潘贤章，王昌昆，等）. Soils(土壤）, 2016, 48(2): 381.
[2] Yin Z, Lei T W, Yan Q H, et al. Computers and Electronics in Agriculture, 2013, 99: 101.
[3] Bowers S A, Hanks R J. Soil Science, 1965, 100(2): 130.
[4] Stoner E R, Baumgardner M F. Soil Science Society of America Journal, 1981, 45: 1161.
[5] Hummel J W, Sudduth K A, Hollinger S E. Computers and Electronics in Agriculture, 2001, 32(2): 149.
[6] Ben-Dor E, Heller D, Chudnovsky A. Soil Science Society of America Journal, 2008, 72(4): 1113.
[7] ViscarraRossel R A, Lobsey C R, Sharman C, et al. Environmental Science & Technology, 2017, 51(10): 5630.
[8] LIU Xian, FEI Liang-jun, WANG Bo, et al(刘显，费良军，王博，等). Journal of Irrigation and Drainage(灌溉排水学报), 2016, 35(9): 20.
[9] WANG Xing-peng, DUAN Ai-wang, LI Shuang(王兴鹏，段爱旺，李双). Transactions of the Chinese Society for Agricultural Machinery(农业机械学报), 2016, 47(6): 100.
[10] Steffens M, Buddenbaum H. Geoderma, 2013, 195: 122.
[11] Burud I, Moni C, Flo A, et al. Soil and Tillage Research, 2016, 155: 523.
[12] LI Shuo, WANG Shan-qin, SHI Zhou(李硕，汪善勤，史舟). Acta Pedologica Sinica(土壤学报)，2015, 52(5): 1014.
[13] YAN Fu-li, WANG Shi-xin, ZHOU Yi, et al(阎福礼，王世新，周艺，等). Journal of Infrared and Millimeter Waves(红外与毫米波学报), 2006, 25(6): 460.
[14] Shi T Z, Chen Y Y, Liu Y L, et al. Journal of Hazardous Materials, 2014, 265: 166.
[15] Wold S, Sjostrom M, Eriksson L. Chemometrics and Intelligent Laboratory Systems, 2001, 58(2): 109.
[16] Suykens J A K, Van Gestel T, De Brabanter J, et al. Least Square Vector Machines. Singapore: World Scientific, 2002.
[17] Wang C K, Pan X Z. Soil Science Society of American Journal, 2016, 80(5): 1393.
[18] Lobell D B, Asner G P. Soil Science Society of American Journal, 2002, 66: 722.
[19] Tian J, Philpot W D. Remote Sensing of Environment, 2015, 169: 280.
[20] ZHANG Jun-hua, JIA Ke-li(张俊华，贾科利). Chinese Journal of Applied Ecology(应用生态学报), 2015, 3: 884.
[21] YAO Yan-min, WEI Na, TANG Peng-qin, et al(姚艳敏，魏娜，唐鹏钦，等). Transactions of the Chinese Society of Agricultural Engineering(农业工程学报), 2011, 27(8): 95.
[22] ZHANG Jun-hua, JIA Ke-li(张俊华，贾科利). Chinese Journal of Applied Ecology(应用生态学报), 2015, 26(3): 884.
[23] Jin X L, Du J, Liu H J, et al. Agricultural and Forest Meteorology, 2016, 218-219: 250.
[24] Jin X L, Song K S, Du J, et al. Agricultural and Forest Meteorology, 2017, 244-245: 57.
[25] LI Chen, ZHANG Guo-wei, ZHOU Zhi-guo, et al(李晨，张国伟，周治国，等). Chinese Journal of Applied Ecology(应用生态学报), 2016, 27(2): 525.