Objective To address the core problems of low efficiency, insufficient data accuracy, and substantial ecological disturbance in the exploration of high-elevation collapse hazards, this study develops an efficient, accurate, and environmentally friendly refined investigation technology system.

Methods Taking the Tianmen Cave collapse in Zhangjiajie as a typical case, we optimized the integration of UAV oblique photogrammetry, terrestrial 3D laser scanning, and close-range photogrammetry, and constructed a "macro–meso–micro" layered complementary data acquisition scheme. The iterative closest point algorithm and Poisson surface reconstruction method were used to achieve high-precision registration and fusion modeling of multi-source data, and to generate a 3D photorealistic point-cloud fusion model with millimeter-level accuracy. Based on this model, the geometric parameters of structural planes in unstable rock masses were intelligently extracted, and the key mechanical parameter (cohesion) was back-calculated from the collapsed mass, and stability calculations and quantitative benefit evaluation were carried out according to relevant specifications.

Results The three-dimensional fusion model with millimeter accuracy in the study area was successfully constructed, and microcracks (width < 1.0 cm) were clearly identified. Based on the cohesion obtained by inversion (C=90 kPa), the stability coefficients of W1-1 and W1-2 that remained uncollapsed were calculated as 1.14 and 1.15, respectively, indicating a critically stable state. Quantitative comparison shows that, compared with traditional manual investigation, the proposed technology system improved fieldwork efficiency by about 40%, reduced the area of vegetation disturbance by about 84%, reduced carbon emissions by about 61%, and improved data extraction efficiency by about 30%.

Conclusion The multi-source data fusion technology system constructed in this study effectively solves the challenges of refined investigation of high and steep terrain and hidden areas, improves the reliability of stability evaluation, and significantly reduces ecological disturbance, forming a scalable green refined investigation solution for high-elevation collapse hazards.