Abstract
Using NIR spectroscopic techniques, soil parameters of raw soil samples collected in the maize, forage and paddy fields were estimated. Soil parameters were soil moisture, SOM, NO3-N, EC and pH. NIR spectral variables were original spectral reflectance, the first derivative, the second derivative and the Kubelka-Munk transform. For soil moisture content, a linear regression model was available in a range of below 30%db or 70-110%db, while an exponential model for a wide range of moisture content up to 110%db. SOM content and EC were estimated with a single spectral regression model. NO3-N content was estimated with a mufti-spectral exponential model and pH with a multi-spectral linear model.
