论文部分内容阅读
通过内蒙古河套灌区义长灌域野外38个测点的光谱试验,提取反射光谱的盐分指数和吸收光谱的特征参数,研究砂土、砂壤土、壤土土壤含盐量的高光谱定量反演技术。结果表明,针对不同土质,提取最佳盐分指数的组成波段和回归建模的精度都不同;壤土提取最佳盐分指数的组成波段为951nm和952nm,回归方程精度最高(R2adjust为0.88);砂壤土提取最佳盐分指数的组成波段为670nm和671nm,回归方程精度最低(R2adjust为0.55);若不考虑土质因素,提取最佳盐分指数的组成波段为651nm和653nm,回归方程的R2adjust为0.58。350~593nm波段范围内,吸收峰位置、不对称性与不同土质土壤含盐量的相关性相对较好,但反演不同土质含盐量的精度一般(R2adjust最大为0.22),这说明可见光-近红外波段范围含盐量光谱的吸收特征不明显。
Through the spectral test of 38 measuring points in the field of irrigated area in Hetao Irrigation Area in Inner Mongolia, the salinity index and absorption spectrum characteristic parameters were extracted to study the hyperspectral quantitative inversion of soil salinity in sandy soil, sandy loam and loamy soil. The results showed that the compositional bands and regression modeling accuracy of extracting the best salt index were different for different soil types. The components of the optimum salt index for loam extraction were 951nm and 952nm, and the regression equation was the highest (R2adjust was 0.88) The compositional bands of extracted optimal salt index were 670nm and 671nm, and the regression equation was the lowest (R2adjust was 0.55). The compositional bands of extracted optimum salt index were 651nm and 653nm without considering the soil quality factors, and the R2adjust of the regression equation was 0.58.350 In the range of ~ 593nm, the correlation between the absorption peak position and the asymmetry is relatively good, but the accuracy of retrieving different soil salt content is generally (R2adjust maximum is 0.22), which shows that visible light - near Absorption characteristics of the spectrum of salt content in the infrared band are not obvious.