Fig. 1. Feature points of part with complex surface. (a) Front view of feature points; (b) isometric side view of feature points; (c) processed feature points

Fig. 2. Element boundary points

Fig. 3. Process of feature point reduction. (a) Elements after separation; (b) boundary points of elements; (c) lower boundary points of elements

Fig. 4. Model of banded feature points

Fig. 5. Schematic of piecewise least square

Fig. 6. Separation process of lower boundary points. (a) Lower boundary points containing partial high points; (b) complete lower boundary points; (c) lower boundary points after filtering

Fig. 7. Trend matching of elements. (a) Eight cases of element trend matching; (b) diagram of element recognition

Fig. 8. Segmentation process for point cloud data. (a) Interior points of surface identified by regional growth; (b) interior and exterior points of enlarged local surface; (c) segmented surface points by combining two methods

Fig. 9. Discrimination process for points near boundary. (a) Magnified local triangular mesh; (b) vector discrimination principle diagram; (c) diagram of distinguishing inner and outer points

Fig. 10. Reconstructed surface and comparison of surface accuracy. (a) Reconstructed surface; (b) comparison of surface accuracy

Fig. 11. Lower boundary extraction process of belt buckle. (a) Feature points of belt buckle part; (b) boundary points of feature points; (c) lower boundary points of feature points

Fig. 12. Comparison of segmentation effect of point cloud data. (a) Point cloud segmented by algorithm in Ref. [4]; (b) point cloud segmented by Geomagic Designx; (c) point cloud segmented by algorithm in this paper

Fig. 13. Reconstructed surface and comparison of surface accuracy. (a) Reconstructed surface; (b) comparison of surface accuracy