土壤重金属铬的高光谱估算模型

Hyperspectral Estimation Model of Heavy Metal Cr in Soil

为探讨利用高光谱遥感快速监测土壤重金属含量的可行性,以石家庄市水源保护区为研究区域,对其土壤光谱数据与重金属铬含量进行相关性分析,运用多元逐步回归、偏最小二乘回归等多种方法建立估算模型。结果表明:土壤光谱特征波段为411、421、425、446、634、894、1 254、1 566、1 909 nm,主要处于土壤铁锰氧化物的特征吸收区域;多光谱变换指标多元逐步回归模型和多光谱变换指标偏最小二乘回归模型的决定系数R^2分别达到0.689和0.668。与单光谱变换指标模型相比,多光谱指标模型估算效果有明显提升。集成多光谱变换指标建模的思路可为土壤重金属铬的遥感监测提供借鉴。

To explore the feasibility of using hyperspectral remote sensing to quickly monitor soil heavy metal content,this paper takes Shijiazhuang water source protection area as the research area,analyzes the correlation between soil spectral data and heavy metal Cr content.Models were established using multiple methods such as multiple linear stepwise regression and partial least square regression.The results show that the spectral characteristics of soil are 411,421,425,446,634,894,1 254,1 566,1 909 nm,which are basically in the characteristic absorption region of soil iron-manganese oxide.The decision coefficients R^2 of the multivariate multiple linear stepwise regression model and the multivariate partial least square regression model are 0.689 and 0.668,respectively.Compared with the univariate model,the estimation effect is significantly improved.The research results can provide reference for the monitoring and estimation of soil heavy metal chromium.