Alleviating Moisture Effects on Remote Sensing Estimation of Crop Residue Cover

Abstract

A robust model for estimating crop residue cover by remote sensing is required to better quantify management of crop residue in agricultural fields. There is a notable lack of robustness in the model, however, when the moisture contents of soil and crop residue vary. To minimize the influence of soil and crop residue moisture on spectral reflectance, we applied a preprocessing method, external parameter orthogonalization (EPO), which estimates the influence of moisture variation on reflectance by performing a principal component analysis on a small set of spectra that represent the variation in moisture. In this study, reflectance spectra (400–2400 nm) for four soils and four crop residues with different moisture conditions were acquired in the laboratory. Reflectance values of mixed scenes with varying proportions of crop residue and soil were simulated using a linear mixture model. The results indicated that moisture significantly affected the cellulose absorption index, the lignin–cellulose absorption index, and the relationships between crop residue cover and these two indices. For applying EPO to remove moisture effects, we proposed two strategies of EPO application that corresponded to two scenarios: a change in moisture in either soil or crop residue, and a change in moisture in both soil and crop residue. The results demonstrated that EPO was able to effectively remove the moisture effects and that the calibration model was more robust after EPO. This study helps to quantify and remove the effects of moisture on the estimation of crop residue cover.