High and steep rock slopes commonly contain numerous discontinuities, while difficulties in close access, direct observation, and accurate measurement hinder reliable instability hazard identification and threaten the safety of major engineering projects. This study proposes an integrated framework for intelligent instability hazard assessment and demonstrates its application on the spillway slopes of the Shuangjiangkou Hydropower Station. The framework includes high-precision three-dimensional slope reconstruction, intelligent rock mass structure interpretation, and automated instability hazard assessment. High-resolution optical imagery and light detection and ranging point cloud data were acquired using a multi-unmanned aerial vehicle collaborative survey strategy and integrated into a slope outcrop fusion model, enabling pixel-level alignment between outcrop textures and point cloud data while accurately reproducing discontinuity geometries. Based on this model, improved extraction algorithms and artificial intelligence-based recognition methods identified 358 planar discontinuities and 597 linear discontinuities automatically. After optimization and false-feature elimination, 533 valid discontinuities were retained. A refined rock mass kinematic analysis method incorporating actual discontinuity geometries and slope curvature parameters accurately located potential instability hazards. Results demonstrate that the proposed framework significantly improves the efficiency, accuracy, reliability, and engineering applicability of instability hazard assessment for high and steep rock slopes in engineering practice.

Slope outcrop fusion model,Rock Mass Structure,Unstable rock mass positioning,Rock slope engineering