A robust 3D point cloud skeleton extraction method

A robust method for three-dimensional point cloud skeleton extraction is proposed in this paper. First, a Laplace operator is used for contracting three-dimensional point cloud locally, and PCA (principal component analysis) is performed on the contracted point cloud to extract skeleton branches. Then, the local points that have been extracted as skeleton branches are fixed while the rest undergo further contraction. Point cloud contraction and local skeleton extraction are repeated until a complete skeleton curve that satisfies the termination conditions is obtained. Finally, by processing the cross points of skeleton and fitting cubic B-spline curves onto them, we can obtain the final skeleton curves. Experimental results show that, compared to existing methods, point cloud skeleton extraction with the proposed method is more robust and resistant to noise.