Lightweight Extraction Method for Structural Feature Detection in Large-Scale Aircraft Panels

In the process of structural feature extraction for large aviation panel structures, conventional methods based on point-based measurement or post-scanning feature extraction measurement suffer from issues like incomplete feature extraction, low efficiency, and insufficient automation. In response to these challenges, this study introduces a lightweight extraction technique tailored for detecting structural features in such components. Firstly, an improved bilateral filtering method, based on normal and curvature information, is applied to smooth each frame of linear point cloud data acquired during the laser scanning process. Subsequently, for the smoothed data, a key inflection point fitting calculation method is proposed to extract critical feature points from each frame of linear point cloud data. After completing feature extraction for all frames of measurement line point cloud data, Euclidean clustering segmentation is employed to partition and preserve key structural features. Finally, experimental validation confirms that the proposed method effectively extracts critical stringers and edge contour features from large-scale panel structures with high precision.