Fig. 1. Occlusion analysis of virtual and real scene fusion display

Fig. 2. Image coding flow based on depth value determination

Fig. 3. Schematic of scenes layering

Fig. 4. Sampling of real and virtual scenes. (a) Single camera timing sampling system；(b) real scene“soldier”model image; (c) virtual scene“BAM！”model image and (d) its depth map

Fig. 5. Test effects of Mask R-CNN on COCO data set

Fig. 6. Test effects of self-made data sets. (a) Image of soldier model; (b) corresponding label image with mask; (c) instance segmentation result

Fig. 7. Three pseudo depth images after black-and-white binary processing

Fig. 8. Part of the encoded perspectives

Fig. 9. Schematic of holographic printing. (a) Schematic of printing optical path; (b) printing sequence of hogels

Fig. 10. Schematic of optical reconstruction system

Fig. 11. Reproductive images in five directions

Fig. 12. Details of experimental result

Fig. 13. Depth comparison between virtual scene and real scene when the camera focus on different positions. (a) Ruler #2; (b) ruler #1