Fig. 1. Layout reconstruction of indoor scene

Fig. 2. Rotation result of scene Livingroom. (a) Before rotation; (b) after rotation

Fig. 3. Point distribution of scene House

Fig. 4. Data projection process

Fig. 5. Line detection result by improved RANSAC algorithm

Fig. 6. Line optimization based on improved mean shift algorithm. (a) Collinear line cluster; (b) optimization result

Fig. 7. Schematic diagram of binary labeling based on ray casting algorithm

Fig. 8. Binary labeling result of scene House

Fig. 9. Results of wall rasterization. (a) Rasterization of single wall; (b) rasterization of complex wall

Fig. 10. Floor and ceiling removal result of scene Livingroom. (a) Original scene; (b) point distribution in Z direction; (c) floor and ceiling removal scene

Fig. 11. Floor and ceiling removal result of scene House. (a) Original scene; (b) point distribution in Z direction; (c) floor and ceiling removal scene

Fig. 12. Preliminary layout division for scene Livingroom. (a) Original point cloud data; (b) data projection result; (c) preliminary layout division result

Fig. 13. Preliminary layout division for scene House. (a) Original point cloud data; (b) data projection result; (c) preliminary layout division result

Fig. 14. Results of layout subdivision. (a) Scene Livingroom; (b) scene House

Fig. 15. Comparison of opening detection result. (a1)--(a3) Real scene; (b1)--(b3) intercept data; (c1)--(c3) detection result obtained by method in Ref. [19]; (d1)--(d3) detection result obtained by proposed method

Fig. 16. Global reconstruction of scene Livingroom. (a) Reconstruction result obtained by three dimensional stacking; (b) global reconstruction result obtained by proposed method; (c) reconstruction result with ceiling removal obtained by proposed method

Fig. 17. Global reconstruction of scene House. (a) Reconstruction result obtained by three dimensional stacking; (b) global reconstruction result obtained by proposed method; (c) reconstruction result with ceiling removal obtained by proposed method