论文部分内容阅读
基于光谱反射率快速、无损的检测优势,以于田地区不同氯化钠含量土壤光谱反射率作为信息源,探讨利用反射光谱估算土壤氯化钠含量的可行性。于2012年7月对于田地区进行野外调查,在测定土壤光谱反射率及氯化钠含量的基础上,对光谱反射率数据进行预处理并建立了逐步多元线性回归(SMLR)模型和偏最小二乘回归(PLSR)模型,并利用决定系数R2和均方根误差RMSE对模型稳定性和预测能力进行检验,进而比较不同预处理方法和不同模型估算结果的适用性。结果表明:反射率二阶微分光谱是预测土壤样本氯化钠含量的最佳光谱指标;偏最小二乘回归(PLSR)模型是建立土壤光谱与氯化钠含量关系的最优模型,R2和RMSE分别为0.812和0.105。利用反射光谱估算土壤氯化钠含量,通过各种光谱预处理方法提高估算精度,可实现在区域尺度上的土壤盐渍化监测和评价。
Based on the fast and non-destructive detection of spectral reflectance, the spectral reflectance of soil with different sodium chloride content in Yutian area was used as the information source, and the feasibility of using soil reflectance spectroscopy to estimate the content of sodium chloride in soil was discussed. In July 2012, a field survey was conducted on the Yutian area. Based on the spectral reflectance and sodium chloride content in the soil, the spectral reflectance data were preprocessed and the stepwise multiple linear regression (SMLR) model and partial least-squares (PLSR) model, and test the model stability and prediction ability by using the determination coefficient R2 and the root mean square error RMSE, then compare the applicability of different pretreatment methods and different model estimation results. The results showed that the second derivative spectrum of reflectance was the best spectral index for predicting the content of sodium chloride in soil samples. Partial least squares regression (PLSR) model was the best model to establish the relationship between soil spectra and sodium chloride content. R2 and RMSE Respectively 0.812 and 0.105. Estimating the sodium chloride content in soil by reflectance spectroscopy can improve the accuracy of estimation by various spectral pretreatment methods and can be used to monitor and evaluate soil salinization at the regional scale.