Discussion on Detection of Sediment Moisture Content at Different Altitudes Employing UAV Hyperspectral Images

This study examined the spectral characteristics of sediments according to moisture content using an unmanned aerial vehicle (UAV)based hyperspectral sensor and evaluated the efficiency of moisture content detection at different flight altitudes.For this purpose, hyperspectral images in the 400-1000nm wavelength range were acquired and analyzed at altitudes of 40m and 80m for sediment samples with various moisture contents.The reflectance of the sediments generally showed a decreasing trend as the moisture content increased.Correlation analysis between moisture content and reflectance showed a strong negative correlation (r < -0.8) across the entire 400-900nm range.The moisture content detection model constructed using the Random Forest technique showed detection accuracies of RMSE 2.6%, R 0.92 at 40m altitude and RMSE 2.2%, R 0.95 at 80m altitude, confirming that the difference in accuracy between altitudes was minimal.Variable importance analysis revealed that the 600-700nm band played a crucial role in moisture content detection.This study is expected to be utilized in efficient sediment moisture management and natural disaster prediction in the field of environmental monitoring in the future.