This study utilized unmanned aerial system imagery to analyze land use on sloping terrain and assessed
the performance of several visible light remote-sensing indices—namely the green leaf index (GLI), normalized difference green-red index (NDGRI), and visible atmospherically resistant index (VARI)—for comparison with the normalized difference vegetation index (NDVI). The study area was located in the mountainous region of Dongshi District,Taichung City, Taiwan. The GLI, NDGRI, and VARI exhibited strong discriminative capabilities for vegetation and bare soil. In a separability index analysis, the vegetation discrimination results of the NDGRI and GLI were similar to those of the NDVI and significantly better than those of the VARI. Moreover, the VARI and NDGRI outperformed the NDVI in the identification of land development area. Overall, the NDGRI emerged as the optimal visible light remotesensing index in this study. The present results could serve as a valuable reference for soil and water conservation in the contexts of ecological conservation, environmental disaster management, land use, and violation enforcement.
Key words: Unmanned aerial system, Remote-sensing index, Land use