| 摘要: |
| 【目的】更好地掌握和利用滨海地区的盐渍土资源，建立高效的光谱模型以监测土壤水分状态。【方法】以黄河三角洲地区滨海盐渍土为研究对象，通过添加不同含盐量的NaCl溶液，模拟自然状态下滨海地区土壤水分的蒸发过程。同时采用近地面高光谱技术测量不同含水率的土壤光谱数据，进行了18种不同的光谱变换，分别建立了土壤光谱数据与土壤含水率之间的偏最小二乘回归模型（PLSR）。【结果】平滑+归一化光谱变换（包含变量归一化、范围归一化、最大值归一化、面积归一化）后的模型均取得了较好的效果，其中平滑+变量归一化后的光谱模型可以直接用于盐渍土水分的反演（R2=0.713 1，RMSE=0.095 0，RPD=1.823 7）。进一步组合不同光谱变换方法之后建模，发现平滑+变量归一化+MSC变换后的模型精度得到了明显的提高（R2=0.866 1，RMSE=0.062 8，RPD=2.764 3）。但是也有部分模型在多种光谱变换方法组合后，精度呈明显下降趋势，说明光谱建模时需要选择适当的光谱变换方法。【结论】本次研究建立的高光谱预测模型稳定性强，预测精度高，所使用的水分数据数值宽泛，能够很好地适用于滨海盐渍土区不同土壤水分状态的遥感监测。 |
| 关键词: 土壤含水率；高光谱模型；光谱变换；滨海盐渍土；人工模拟 |
| DOI：10.13522/j.cnki.ggps.2019280 |
| 分类号: |
| 基金项目: |
|
| Using Hyperspectral Imagery to Estimate Soil Moisture and Calculate Evapotranspiration from Coastal Saline Soil |
|
ZHANG Xiaoguang， KONG Fanchang
|
|
1. College of Resources and Environment, Qingdao Agricultural University, Qingdao 266109, China;2. State Key Laboratory of Soil and Sustainable Agriculture, Nanjing 210008, China
|
| Abstract: |
| 【Objective】Soil moisture is a key factor mediating water and salt movement in soil. Monitoring soil moisture dynamics is hence critical to ameliorating soil salinity and safeguarding crop production, especially in coastal areas. The purpose of this paper is to test the feasibility of using spectral imaging techniques to estimate soil water content.【Method】The field experiment was conducted at a coastal saline soil in the Yellow River delta. Water with different solute concentrations was applied to the soil first, and we then measured its effect on water evaporation. The spectral imageries of the soil under different water content and salt content were acquired using a surface hyperspectral remote sensing apparatus. Eighteen spectral transformations were used to process the spectral data, and the partial least square regression (PLSR) model between soil moisture content and the spectral data was derived for each transformation. 【Result】Among all transformations, the models based on smoothing+normalization spectral transformation (contained normalization of variables, range normalization, maximum normalization, and area normalization) gave good results. The model based on smoothing + normalization of variables was most accurate with R2=0.713 1 and RMSE=0.095; its ratio of standard deviation to root mean square error (RPD) was 1.823 7, indicating that this model can be directly used to inversely estimate soil moisture of the saline soil. While different combinations can be constructed from the spectral transformations, the combination derived from post-transformation of the smoothing+ normalization of variables+ multiplicative scatter correction model worked best, with R2=0.866 1, RMSE=0.866 1 and RPD=2.764 3. The models derived from other combinations of the spectral transformation were less accurate, indicating that care needs to be taken in selecting the suitable spectral transformation to estimate soil moisture. 【Conclusion】The model presented in this paper was stable and accurate, and its soil moisture estimates can be used to calculate soil evaporation at a wide range of circumstances. The spectral imaging technique can thus be used to real-time monitor soil moisture change in saline soils in coastal areas. |
| Key words: soil moisture; hyperspectral imagery; spectral transformation; coastal saline soil; artificial simulation |
