A SLAM-based high-resolution full-character debris-flow channel morphological mapping system

Enhancing the study of channelized debris flow necessitates precise and high-resolution mapping of channel topography and deposit conditions. Existing mapping technologies like satellite imaging and drone photogrammetry face challenges in accurately observing the interiors of extensive mountainous gullies, particularly in regions affected by events like the Wenchuan Earthquake. Despite the emergence of Simultaneous Localization and Mapping (SLAM) as a 3D mapping technology, its efficacy is hampered in rugged gullies due to two primary challenges: (1) Unusual terrain features and (2) Severe sensor swaying and oscillation, causing significant deviations and noise in SLAM-generated results. Addressing these challenges, we propose an innovative SLAM-based debris-flow channel detection and mapping system. It incorporates three key enhancements to refine SLAM outcomes: (1) A deviation correction algorithm assisted by digital orthophoto maps effectively mitigates systematic errors; (2) A point cloud smoothing algorithm significantly reduces noise levels; and (3) A cross-section extraction algorithm enables quantitative assessment of channel deposits and alterations. Conducting field experiments in Chutou Gully, Wenchuan County, China, in February and November 2023—representing pre and post-rainy season observations—validated the system's capabilities in markedly improving SLAM results. This advancement facilitates SLAM's efficacy in mapping challenging terrains, compensating for existing technology limitations in detecting debris flow channel interiors. The system aids in delineating detailed channel morphology, erosion patterns, deposit differentiation, volume estimation, and change detection. By rectifying the shortcomings of current methodologies, this methodological approach serves to augment the understanding of full-scale debris flow mechanisms, long-term post-seismic evolution, and hazard assessment in affected regions.