Real-Time Analysis of Soil Total Nitrogen and Soil Total Phosphorus with NIR Spectroscopy

The grey-brown alluvial soil is a typical soil in the Northern China. It is selected as research object to reveal feasibility and possibility of real-time analyzing soil parameter with NIR spectroscopic techniques. 150 samples are collected from a winter wheat farm. And then the NIR absorbance spectra are rapidly measured under the original conditions by a Nicolet Antaris FT-NIR analyzer. Two soil parameters, soil total nitrogen content (TN) and soil total phosphorus content (TP), are analyzed. 2262 nm and 1888 nm wavelengths were selected in the TN estimated model. Estimation model of soil TN content by using the first spectral derivation was established. After noises was eliminated by the first spectral deviation, the correlation coefficient of the model (R-C(2)) was 0.767. For TP, a partial least square regression model was available, using 2258 nm, 2040 nm, 1810 nm, 1490 nm, 1332 nm and 996 nm wavelengths with correlation coefficient of 0.818. The results showed that soil total nitrogen content and soil total phosphorus content could be evaluated by using NIR absorbance spectra of soil samples.