基于地貌类型的土壤有机质多光谱遥感反演

Multi-spectral Remote Sensing Inversion of Soil Organic Based on Landform Types

基于地貌类型分析土壤有机质含量与多光谱遥感影像光谱波段之间相关关系,构建不同地貌类型区有机质含量反演模型。结果表明,各波段光谱反射率与土壤有机质含量均呈负相关关系。利用SPSS软件对所有波段进行剔除变量(remove)线性回归分析,当全部波段参与构建反演模型时,一次反演模型拟合效果较好。分地貌类型区构建土壤有机质反演模型精度高于整个区域反演模型精度,与实际值对比,当允许误差为7%时,土壤有机质含量识别度为91.65%。基于地貌类型构建土壤有机质含量反演模型提取研究区土壤有机质含量切实可行,且精度较高。

Based on landform types,the correlation of soil organic and multi-spectral band of remote sensing image was analyzed in the paper,and the inversion models of soil organic were built.The result showed that the correlation of spectral reflectance and the content of soil organic matter were negative.Using SPSS,linear regression analysis of remove were done to all the bands.An inversion model had the better fitting effect while all the bands were used to build inversion models.The soil organic models based on landform types had the higher precision than the model of all the sampling points.Compared with the actual value,it had the identification of degree 91.65%of the allowed error 7%.The conclusion is that it has the higher feasibility and precision of using inversion models to extract soil organic matter based on landform types.