土壤钾含量高光谱定量反演研究

Quantitative Inversion of Soil Potassium Content by Using Hyperspectral Reflectance

为了更快捷准确地进行土壤钾(K)含量的预测,基于土壤高光谱数据和化学元素分析数据,研究土壤光谱与土壤钾含量之间的定量关系。在对土壤原始光谱进行处理分析基础上,提取反射率(R)、反射率倒数的对数(log(1/R))、反射率一阶微分(R')和波段深度(BD)4种光谱指标,运用偏最小二乘回归方法建立相应的预测模型,并对模型进行检验。结果表明,波段深度是估算土壤钾含量最好的光谱指标,其建模精度超过0.85,均方根误差不超过0.1;全波段高光谱分辨率反射光谱具有快速有效估算土壤钾含量的潜力。

In order to predict the soil potassium content more quickly and accurately,this paper studied the relationship between soil spectrum and soil potassium content based on soil hyper-spectrum and chemical element analytical results.Based on preprocessing,the authors extracted the original soil spectrum and four parameters of spectra,i.e.,reflectance spectra,first derivative reflectance spectra,inverse-log spectra and band depth,so as to establish the prediction model for potassium content by PLSR.The model was tested,and the results indicate that band depth is the optimum parameter for inverting soil potassium content,with a minimum modeling correlation coefficient of 0.85 and maximum RMSE of 0.1.This research shows that,as a non-destructive method,the soil spectrum with high spectral resolution in the whole range has the potential for the rapid simultaneous prediction of potassium concentration.