Field reconnaissance has difficulty in providing information on topographic features of debris flow in vegetation-covered mountain regions. The study used drone-captured multispectral imaging, digital terrain models, and geomorphic indices to investigate the sources and erosion patterns of gravelly debris flows in mountainous areas. Geomorphic indices, including the topographic wetness index (TWI), stream power index (SPI), sediment transport index, terrain ruggedness index (TRI), and the vegetation index normal differential water index (NDWI), were used to identify key topographic and hydraulic characteristics that contribute to the initiation of gravelly debris flows. The analysis revealed that the debris flows originated from the softening of yellow soil and the dislodging of gravel in loess areas, both induced by surficial water infiltration. Areas with high TWI and SPI values were vulnerable to erosion from surficial water, leading to the formation of erosion gullies. Likewise, regions with high TWI and NDWI values were likely starting points for upstream landslides. Conversely, areas characterized by low TWI and TRI values but high NDWI were prone to downstream sediment deposition. Drone-based multispectral imaging, augmented by geomorphic and vegetation indices, effectively captures the characteristics of vegetation-covered areas, thereby enhancing debris flow field investigations in an inaccessible mountain area.

野外勘察难以提供植被覆盖山区泥石流地形特征信息。该研究利用无人机捕获的多光谱成像、数字地形模型和地貌指数来研究山区砾石泥石流的来源和侵蚀模式。地貌指数，包括地形湿度指数（TWI）、水流功率指数（SPI）、泥沙输送指数、地形崎岖指数（TRI）和植被指数正态差分水指数（NDWI），用于识别关键地形和水力指数。有助于引发砾石泥石流的特征。分析表明，泥石流源于黄土软化和黄土地区砾石脱落，两者都是由地表水入渗引起的。 TWI和SPI值较高的地区容易受到地表水的侵蚀，导致侵蚀沟的形成。同样，TWI 和 NDWI 值较高的区域可能是上游滑坡的起点。相反，TWI 和 TRI 值低但 NDWI 值高的区域容易发生下游沉积物沉积。基于无人机的多光谱成像，辅以地貌和植被指数，可有效捕捉植被覆盖区域的特征，从而加强难以进入山区的泥石流现场调查。