【摘要】 [目的]为科学和合理种植作物以及提高作物产量提供参考数据。[方法]采集研究区48个点的盐渍地土壤,测得土壤含水量数据,并且在自然光源和人工光源条件下采集土壤光谱反射率数据,对反盐期不同深度土壤的含水量进行统计分析,研究土壤烘干前和烘干后的光谱反射率,对自然光源的土壤光谱反射率进行一阶、二阶、三阶、四阶、五阶和六阶导数微分处理,剖析人工光源的土壤光谱反射率数据、自然光源的光谱反射率、导数微分变换后的光谱反射率数据与土壤含水量之间的相关分析,进行一元一次、一元二次、指数、二元逐步回归、偏最小二乘回归(PLSR)函数的拟合,并用总均方根误差(RMSE)以及显著性水平验证拟合函数的精度。[结果]表层土壤(0~10 cm)的含水量明显低于10~30 cm深度的土壤含水量,且表层土壤含水量由于土壤盐渍化程度的不同而存在差异,绿洲内部和外围的土壤含水量也存在明显差异;干燥土壤的光谱反射率曲线比湿润土壤更高;人工光源、自然光源获取的土壤光谱反射率数据与土壤含水量之间呈负相关关系,人工光源的相关性比自然光源的相关性略高;自然光源获取的光谱反射率经过导数微分形式的变换后与土壤含水量之间的相关性得到很好的提高;人工光源的最佳拟合方程为一元二次方程Y=266.4X21680+35.2X1680+1.9,RMSE为0.33,显著性水平P<0.01;自然光源的最佳拟合方程为偏最小二乘方法的光谱反射率的四阶形式,RMSE为0.89,显著性水平P<0.01。[结论]该研究能够为干旱区土壤含水量和光谱特征的关系的研究提供数据支撑,奠定高光谱技术在干旱区的应用基础。更多还原

【Abstract】 [Objective] The research aimed to provide the reference data for the scientific and regular plantation of the crop and the improvement of the crop yield.[Method]Saline soils in 48 points were collected in the study area,and the soil moisture data and soil spectral reflectance data in natural light and artificial light conditions were get to analyze moisture content of the soil at different depth statistically.Spectral reflectances before and after soil drying were studied.Differentiation processing of first-order,second-order,third-order,fourth-order,fifth-order and sixth-order derivative to soil spectral reflectance of natural light was used.Spectral reflectance data of artificial light and natural light soil were analyzed.Derivative differential conversion between spectral reflectance data and soil moisture was analyed,fitting various functions,and verified.[Result] The moisture of topsoil(0-10 cm) was significantly lower than that of the deeper soil.It decreased as the distance away from basin and the degree of salinity increased.Dry soil spectral reflectance was higher than the spectral reflectance curve of moist soil.Whether it was artificial light or natural light condition,there were a negative correlation between the spectral reflectance data and soil moisture,but the correlation of artificial light was slightly higher than the natural light.The best fitting function of artificial light source was quadratic equation with one unknown while RMSE was 0.33,and P < 0.01,and the best fitting function of natural illumination was partial least squares while RMSE was 0.89,and P < 0.01.[Conclusion] Accurate estimation of soil moisture content and its spectral features could provide a reference for planting crops scientifically and reasonably,and it could lay the foundation of hyperspectral data application.更多还原