Estimation of soil moisture with hyperspectral reflectance based on ground-based remote sensing

Abstract

A lab experiment and a field experiment were conducted to develop a method to estimate soil moisture level using a handheld spectroradiometer. In lab experiment, forty-one soil samples with different moisture content were prepared. The hyperspectral reflectance of all soil samples were measured, and the moisture content of all soil samples were analyzed. In field experiment, a 50×300 m 2 small farm was selected, divided into 150 10×10 m 2 grids, and the hyperspectral reflectance of each grid was measured. A soil sample (about 1000 g) was collected for subsequent moisture analysis from each grid. The results from lab experiment showed that there was a linear correlation between soil moisture content and soil hyperspectral reflectance when the moisture content was below 19 %db. By comparing the two experiments, we also identified that the soil surface condition and the surrounding temperature can influence the precision of the hyperspectral reflectance measured in the field experiment. Based on the field experiment, several hyperspectral reflectance maps were drawn and they show high correlation with soil moisture content map. All results from both the lab experiment and the field experiment show possibility to estimate soil moisture content with hyperspectral reflectance obtained by ground-based remote sensing.