水分对土壤有机质检测影响的光谱特性分析及抗水分干扰模型建立

Analysis of the Effect of Moisture on Soil Organic Matter Determination and Anti-Moisture Interference Model Building Based on Vis-NIR Spectral Technology

土壤水分对光谱表现出很强的吸收性,且土壤水分与土壤有机质的吸收波段有重叠,因此土壤水分对土壤有机质的检测造成一定的干扰。为此做了以下工作：（1）采用可见近红外光谱仪在室内获取相同含水率下不同土壤动态光谱图;（2）通过对相同含水率下不同有机质含量的二维同步相关光谱图分析得出：当土壤为烘干土样时,600和1 660nm左右表征土壤有机质的波段出现强的自相关峰,但随着含水率的增加,这两个波段逐渐消失,由于受水分的影响,1 931,2 200和1 480nm均形成了强的自相关峰。说明水分会掩盖表征土壤有机质信息的波段,对土壤有机质检测造成干扰。（3）为了消除水分影响,提高模型对不同含水率下土壤有机质的预测精度,将田间近似最大含水率样本参与建模,采用偏最小二乘定量分析方法在550-650和1 610-1 710nm波段内建立了抗水分干扰土壤有机质预测模型,并对不同含水率的土壤有机质进行预测,结果表明：预测样本的相关系数为0.954,标准偏差为0.744%,标准差为0.844%,预测效果明显提高,说明此方法可减少水分对土壤有机质检测的影响。

Soil moisture shows a strong absorption for spectroscopy while soil organic matter and moisture have identical absorption bands.Therefore,the soil moisture causes interference to soil organic matter detection.The study made the following works：firstly,different soil organic matter dynamic spectrums under different moisture content were acquired with visible near-infrared spectroscopy;secondly,different organic matter content features under the same moisture content were analyzed with two-dimensional synchronization correlation spectroscopy.When the soil moisture is 0%,around 600 and 1 660 nm which characterize soil organic matter band appear strong autocorrelation peaks.With moisture content increasing,around 600 and 1 660 nm band disappear and around 1 931,2 200 and 1 480 nm band appear strong autocorrelation peaks in the near-infrared region.Soil moisture covers information bands which characterize soil organic matter and affects soil organic matter detection;thirdly,the maximum moisture content samples approximately in the filed participated modeling to eliminate the effect of moisture on soil organic matter detection and improve the model prediction accuracy.The anti-moisture interference prediction model which used550-650and 1 610-1 710 nm wavelengths by PLS（Partial Least Squares）quantitative analysis method was established to predicate the soil organic matter content under different moisture content.The results are as follows：predicted correlation coefficient,SEP and RMSEP is 0.954,0.744 and 0.844 respectively.Predicted effect improves significantly.It is that the method can reduce the effect of moisture on soil organic matter detection.