论文部分内容阅读
为实时、准确地获取原位土壤含水量信息,利用可见/近红外光谱技术,分别使用全局偏最小二乘(PLS)建模、局部PLS建模方法,对田间原位土壤含水量进行快速估测。结果表明:全局PLS模型中,其建模集的决定系数(R~2)、交叉验证均方根误差(RMSECV)分别为0.943和1.750%,检验集的决定系数(R~2)、预测均方根误差(RMSEP)分别为0.956和1.260%。局部PLS模型中,分别比较了选取定标子集的2种方法(欧氏距离法和马氏距离法),采用欧氏距离法和马氏距离法选取定标子集进行建模的R~2值分别为0.974和0.979,RMSEP值分别为0.976%和0.943%。因此,将可见/近红外光谱技术应用到田间原位含水量测量是可行的,其中,使用局部建模方法的效果优于全局建模。
In order to obtain the in situ soil water content information in real time and accurately, the PLS modeling method and the PLS modeling method were used to estimate the in situ soil moisture content by using visible / near infrared spectroscopy Measurement. The results show that the coefficient of determination (R ~ 2) and the root mean square error of cross validation (RMSECV) are 0.943 and 1.750% respectively in the global PLS model. The coefficient of determination (R ~ 2) Root mean square error (RMSEP) was 0.956 and 1.260% respectively. In the local PLS model, two methods (Euclidean distance method and Mahalanobis distance method) for selecting the scaled subset were compared respectively. The Euclidean distance method and Mahalanobis distance method were used to select the calibration subset for modeling R ~ 2 values were 0.974 and 0.979, RMSEP values were 0.976% and 0.943% respectively. Therefore, it is feasible to apply visible / near-infrared spectroscopy to field measurements of in situ water content, and the effect of using local modeling method is better than that of global modeling.