Hyperspectral Bare Soil Index (HBSI): Mapping Soil Using an Ensemble of Spectral Indices in Machine Learning Environment

Spectral remote-sensing indices based on visible, NIR, and SWIR wavelengths are useful in predicting spatial patterns of bare soil. However, identifying an effective combination informative or spectral for mapping soil a complex urban/agricultural region is still challenge. In this study, we developed new bare-soil index, the Hyperspectral Bare Soil Index (HBSI), to improve accuracy mapping. We tested HBSI using high-spectral-resolution AVIRIS-NG Sentinel-2 multispectral images. applied ensemble modeling approach, consisting random forest (RF) support vector machine (SVM), classify found that outperformed other existing with over 91% AVIRIS-NG. Furthermore, normalized difference vegetation index (NDVI) showed better performance classification, >92% >97% Also, RF-SVM surpassed individual models. The novelty due its development, since it utilizes blue band addition NIR SWIR2 bands from data