Fig. 1. Gray distribution of speckle feature points

Fig. 2. Speckle descriptor establishment process. (a) Window size selection; (b) descriptor point set

Fig. 3. Descriptor similarity calculation. (a) Descriptor matrix $K_p$ of point p; (b) descriptor matrix $K_q$ of point q; (c) $K_{\mathrm{x}\mathrm{o}\mathrm{r}}$ result

Fig. 4. Flow chart of speckle feature point propagation matching

Fig. 5. Experimental equipment

Fig. 6. Speckle feature point extraction effect. (a) Speckle images; (b) method in Ref.[3]; (c) method in Ref. [11]; (d) method in Ref. [4]; (e) proposed method

Fig. 7. Experimental results of speckle feature point extraction. (a) Number of speckle feature points; (b) extraction accuracy

Fig. 8. Effect of descriptor size on matching results. (a) Matching quantity; (b) mismatching rate

Fig. 9. Experimental results of different speckle feature point matching methods. (a) Matching quantity; (b) mismatching rate

Fig. 10. 3D reconstruction results of different matching algorithms. (a) Method in Ref.[8]; (b) method in Ref.[13]; (c) proposed method

Fig. 11. Effect of speckle feature point extraction methods on matching results. (a) Matching quantity; (b) mismatching rate

Fig. 12. 3D reconstruction results of different feature point extraction methods. (a) Method in Ref.[11]; (b) method in Ref. [4]; (c) proposed method