摘要：Hyperspectral data provide valuable information for salt content estimation. In this paper, soil sampleswere collected from the Keriya Oasis, Southern Xinjiang, China. Samples were bagged, brought to thelaboratory, air-dried, ground, and sieved using 2 mm size sieve. Soil salt contents were measured andthe reflectance spectra were collected using FieldSpec3 in laboratory condition. The continuum removal(CR) reflectance was obtained after smoothing and averaged spectral data conversion of 10 nminterval. A total of 8 spectral slopes at the wavelength between 365-375 nm, 1435-1465 nm, 1855-1865 nm, 1915-1925 nm, 2085-2095 nm, 2295-2315 nm, 2365-2395 nm and 2465-2475 nm werecalculated based on the correlation analysis between soil salt content and its spectrum. Thirty of 40samples were used for establishing hyperspectral model for estimating soil salt content and the other10 samples were for the model verification. The multiple linear regression (MLR) and partial leastsquares regression (PLSR) were used to model and estimate soil salt content. The results showedthat, when soil salt content is higher than 2.10 g·kg -1 , spectral slope values increase with the increaseof salt content. The estimation accuracy of the model based on MLR was higher than the model basedon PLSR. The R 2 for calibration and validation of the optimum multiple linear regression model were upto 0.834 and 0.664, respectively, and its RMSE values of calibration and validation were 2.9707 and3.2691, and the RPD value was 2.09, respectively. This spectral slope based model was asupplementary modelling for hyperspectral soil salinity estimation, and can be a basis for futuresatellite-based hyperspectral monitoring and evaluation of soil salinity.
关键词：Soil salt content;
Spectral slope;
Hyperspectral reflectance;
Keriya Oasis
