Study on the Pretreatment of Soil Hyperspectral and Na+ Ion Data under Different Degrees of Human Activity Stress by Fractional-Order Derivatives

Soluble salts in saline soil often exist the form of salt base ions, and excessive water-soluble ions can harm plant growth. As one Na+ ion, is main indicator degree salinization. The pretreatment visible, near infrared short-wave (VNIR-SWIR) spectroscopy data key to establishing a high-precision inversion model, proper method fully extract effective information hidden hyperspectral data. Meanwhile, different degrees human activity stress will have an impact on ecological environment oases. However, there are few comparative analyses effects for under interference conditions. Therefore, this study, difference disturbance caused by activities was used as basis dividing experimental area into lightly disturbed (Area A), moderately B) severely zone C). Grünwald-Letnikov fractional-order derivative (FOD) preprocess VNIR-SWIR spectroscopic measured FieldSpec®3Hi-Res spectrometer, which could useful FOD results avoid loss traditional integer-order (1.0-order, 2.0-order) pretreatment. spectrum composed five transform spectra (R, R, 1/R, lgR, 1/lgR) 21 methods (step size 0.1, range from 0.0- 2.0-order). In addition, manuscript compares analyzes advantages between integer-order. were follows: (1) reveal nonlinear characteristics variation laws field soil, namely, due continuous performance order selection, accurately depicts details spectral changes during derivation process, improves resolution peaks spectrum. (2) There big shape correlation coefficient curve original at FOD. has clear outline rang 0.6-order, change trend gentle, presents certain gradual form. With increase FOD, gradually increased, fluctuation greater 1.0-order 2.0-order. (3) Regardless transformation regions, improvement effect significantly better than that derivative. Comparative analysis shows he percentage former more 3%, highest 17%.