基于实验室高光谱数据的大围山森林土壤氧化铁全量反演

氧化铁是土壤中含铁矿物的主体,是土壤发育和土壤分类最明显和最有用的指标之一。本文以湖南省大围山森林土壤为研究对象,通过实验室化学成分测定和光谱采集,在光谱预处理及组合变换基础上,采用相关性分析筛选土壤氧化铁全量的敏感波段,并分别建立多元逐步回归和偏最小二乘回归反演模型。结果表明:不同土壤光谱曲线趋势基本一致,均形似陡坎,且在420～580 nm波段,土壤氧化铁全量与光谱反射率呈负相关关系;不同的光谱数据变换方式可以提高光谱与氧化铁全量的相关性,Savitzky-Golay(S-G)平滑和去包络线相结合优于其他预处理方法;土壤氧化铁全量的特征波段主要为392、427、529、523、549、559、565、570、994和1040nm,偏最小二乘回归模型比多元逐步回归模型具有更好的稳定性,适合于快速估算红黄壤区森林土壤氧化铁全量。

Inversion of Iron Oxide Contents in Forest Soils of Dawei Mountains Using Laboratory Hyperspectral Data

Iron oxide is the main body of iron-bearing minerals in the soil and is one of the most obvious and useful indicators of soil development and soil classification, and its content has an important impact on mountain ecology and vegetation growth. To further explore the ability of hyperspectral techniques to estimate soil trace element content and the optimal spectral pretreatment of these elements, forest soils of Dawei Mountain, Hunan Province, were selected for the study. Based on the spectral pretreatment and combinatorial transformation, the sensitive bands of total iron content were selected by correlation analysis, and multiple stepwise regression and partial least squares regression models were constructed. The results showed that: (i) the spectral curves of different soil total iron content had the same trend in the whole band range, and all of them were shaped like steep hills; in the band from 420 nm to 580 nm, the iron oxide content was negatively correlated with the spectral reflectance; (ii) different spectral data transformation methods could improve the correlation between spectra and soil iron oxide, and the combination of Savitzky-Golay smoothing and de-enveloping lines was better than others; (iii) Hyperspectral combined with multiple stepwise regression and partial least squares regression models can predict the total iron content in red and yellow soils quickly and accurately. In particular, the partial least squares regression model has good stability and is suitable for the rapid estimation of total iron content in forest soils in red and yellow loam areas.