In order to explore the quantitative method of quantitative determination of heavy metal Cd content in chestnut soil,the hyper-spectral data and Cd content of 60 Chestnut soil samples in Baiyunebo area of Inner Mongolia were measured.The sample data were re-sampled at different intervals.The reciprocal logarithm,the depth of the band and the first order differential were calculated based on the spectral reflectance of the original sample and the reflectance of the soil sample after resampling.Then the soil samples were divid-ed into two parts:one was used to establish the inversion model of the characteristic bands of reflectance and different spectral transform pattern and the heavy metal Cd content by a partial least -square method,and the other was used for model validation.The results show that quantitative inversion of heavy metal Cd content in chestnut soil is practical;the transform of soil spectral data can effectively im-prove the stability and prediction ability of the inversion model;when modeling spectral indices are the same,the model's stability and predictability are not proportional to the resolution of the modeling spectra.%为了探索定量反演内蒙古中西部栗钙土重金属Cd含量的最佳方法,本文采集并测量了内蒙古白云鄂博矿区60份栗钙土样品的高光谱数据和重金属Cd含量,按照不同波段间隔对样品数据进行重采样,并对原始光谱分辨率和重采样后土壤样品的反射率求取倒数对数、波段深度和一阶微分3种光谱指标.将土壤样品分成两份,一份利用偏最小二乘法建立基于不同光谱分辨率和不同光谱指标的重金属Cd含量反演模型,一份用于模型的验证.结果表明,栗钙土重金属Cd含量定量反演切实可行;土壤光谱数据进行光谱指标的变换,可以有效提高反演模型的稳定性和预测能力;当建模光谱指标相同时,模型的稳定性和预测能力并不与建模光谱分辨率呈正比例关系.
展开▼
