Fig. 1. Process of aircraft skin registration

Fig. 2. Feature points clustering based on Euclidean distance

Fig. 3. Establishment of FPFH local coordinate system. (a) Influence domain of P₀; (b) local coordinate system

Fig. 4. Flow chart of point cloud registration algorithm based on Cκ key point extraction

Fig. 5. Feature point extraction results of Bunny and Dragon. (a) Original point cloud of Bunny; (b) feature points of Bunny in view 1; (c) feature points of Bunny in view 2; (d) original point cloud of Dragon; (e) feature points of Dragon in view 1; (f) feature points of Dragon in view 2

Fig. 6. Registration results of Bunny and Dragon. (a) Initial registration result of Bunny; (b) initial registration result of Dragon; (c) accurate registration result of Bunny; (d) accurate registration result of Dragon

Fig. 7. Aircraft skin point cloud and CAD point cloud. (a) Three-dimensional display of point cloud; (b) three-dimensional display of point cloud in software

Fig. 8. Contour of aircraft skin. (a) Three-dimensional display of skin contour; (b) three-dimensional display of skin contour in software

Fig. 9. Selection of feature point. (a) Curve between threshold and characteristic area number; (b) extraction effect of feature point

Fig. 10. Registration results of aircraft skin. (a) Initial registration result; (b) accurate registration result

Fig. 11. Overall registration results of aircraft skin

Fig. 12. Comparison of registration results of each method. (a) Registration contour effect of FPFH-SAC-IA algorithm; (b) overall effect of FPFH-SAC-IA algorithm;(c) registration contour effect of ICP algorithm;(d) overall effect of ICP algorithm;(e) registration contour effect of proposed algorithm; (f) overall effect of proposed algorithm

Table 1. Comparison of initial registration results