Precision UAV image-to-BIM registration through contour-based matching for concrete dam structural health monitoring

The accurate mapping of UAV images to BIM is critical for long-term structural health monitoring. However, the complexity of concrete dams introduce positioning deviations, making it challenging to precisely localize defects. This paper proposes a precise image-to-BIM method based on contour matching. After point cloud registration establishes global coordinate transformation, the structural contour templates are extracted for UAV viewpoints. The concrete dam contours are classified as outer or inner, and the UAV images are matched with templates using different algorithms. 2D pixel variations are then converted into 3D spatial displacements, and the UAV coordinates are iteratively refined for accurate contour alignment. The corrected coordinates map the detected defects onto the BIM surface to ensure precise image-to-BIM registration. The proposed method is validated on a real concrete dam. After coordinate correction, cracks detected in the UAV images are accurately mapped to the BIM surface, with consistent localization across different viewing angles.