EFFECT OF FARM MACHINERY ON SOME SOIL PROPERTIES USING HYPERSPECTRAL DATA

Abstract

ABSTRACT Traditional soil analyses are expensive, time-consuming and may also result in environmental pollutants. The objective of this study was to develop a methodology to predict soil physical properties specifically the soil bulk density using spectral signature as an alternative to traditional methods. In addition to critically examine the suitability of Visible and Near-infrared (Vis- NIR) portion of the electromagnetic spectrum (350–2500 nm) measurements for calibration procedures and methods of predicting the value of soil bulk density under five different seedbed preparation systems. Spectroradiometer ASD measured the soil spectral reflectance of each system. Stepwise multiple linear regression (SMLR) was used to construct calibration models subjected to the independent validation by SPSS program. The obtained regression models were of good quality (R = 0.959, 0.868, 0.811, 0.751 and 0.677 for system 5, system 1, system 3, system 4 and system 2, respectively). Thus, Visible and Near-infrared (Vis-NIR) reflection spectroradiometer is cost- and time-effective procedure that can be used as an alternative to the traditional methods of measuring of soil physical properties. 2