土壤性质指标光谱反演数据变换模型研究

Data transformation model analysis on spectral retrieved deduction of soil properties index

本文采用土壤含铁量和有机质含量作为性质指标,对反射率作常规模型变换,变换后的数值与性质指标的相关系数最大值有一定提高。又进行反射率一阶微分乘以平方根一阶微分等几种创新模型变换,以及将变换效果较好的个别波段选出进行交叉组合模型变换,变换后的数值与性质指标的相关系数最大值又有一定提高。实验证明,反射率一阶微分乘以平方根一阶微分模型以及选出波段交叉组合变换模型可以有效提高反射率与土壤性质指标的相关性。

Using soil Fe and soil organic matter as properties index in this paper, ordinary types of transformation were applied to the original reflection rate, and the correlation coefficients between the transformed value and soil properties index were calculated, the maximum value increased greatly. Other types of new transformation such as the order 1 derivative of the logarithm times order 1 deriva- tive of the square root were applied, or selecting wavebands which could get bigger correlation coefficients after transformation to take cross-combination, the maximum correlation coefficient between the transformed value and soil properties index got another increase. The results showed that the transformation of the order 1 derivative of the logarithm times order 1 derivative of the square root or cross- combination could make the correlations increase effectively.