本文利用实验室实测的土壤反射光谱以及铅、镉、汞等重金属元素数据,进行土壤重金属元素含量快速预测的可行性研究。本文利用偏最小二乘回归方法,研究了反射率(R)、一阶微分(FDR)、反射率倒数的对数(lg(1/R))和波段深度(BD)等对预测精度的影响,对这几种光谱指标预测土壤重金属含量的能力进行了分析和评价,同时分析了多光谱数据估算土壤重金属元素含量的可行性。结果表明,反射率倒数的对数lg(1/R)是估算土壤重金属元素含量最好的光谱指标,尤其是Cd和Pb,检验精度R超过0.82。有机质、铁锰氧化物和黏土矿物对土壤重金属元素的吸附是可见光—近红外—短波红外光谱估算其含量的机理。多光谱数据同样具有估算土壤重金属元素含量的能力,但实际数据则要考虑多种因素的影响。

This paper analyzed the possibility of reflectance spectra obtained under laboratory conditions for the assessment of Pb,Cd and Hg content in soil quickly.Besides original spectra(R),several spectral indices were also calculated: first derivative reflectance spectra(FDR),inverse-log spectra(lg(1/R)) and band depth(BD).Partial least square regression(PLSR) was used to develop calibrations between spectral indices data and content of soil elements.Coefficient of determination(R) and root-mean-square error(RMSE) were used as criteria for best model.The results show that: 1) lg(1/R) is the best index to estimate soil heavy metal content,especially for Cd(R=0.8221) and Pb(R=0.8612);2) The mechanism for estimating soil heavy metal element content by VIS-NIR-SWIR spectra is the absorption function of organic matter,iron-manganese oxide,and clay minerals;3) Simulated multi-spectral data have the good ability to estimate soil heavy metal elements content.While satisfactory results are obtained by laboratory spectra,more factors should be considered when using field data even satellite data.