Wenjian Li, Shaoyan Gai, Jian Yu, Feipeng Da. Absolute Phase Recovery of Single Frame Composite Image Based on Convolutional Neural Network[J]. Acta Optica Sinica, 2021, 41(23): 2312001

Search by keywords or author
- Acta Optica Sinica
- Vol. 41, Issue 23, 2312001 (2021)

Fig. 1. Principle diagram of absolute phase is obtained by using phase shift method combined with gray code method

Fig. 2. Design of composite fringe pattern. (a) Generation principle of composite fringe pattern; (b) composite fringe pattern

Fig. 3. Principle of proposed method

Fig. 4. Structure of CNN 1

Fig. 5. Structure of residual block network

Fig. 6. Structure of CNN 2

Fig. 7. Partial training data of CNN

Fig. 8. Loss function curves of CNN 1 and CNN 2 under different conditions. (a) Loss curves of molecular term M in CNN 1; (b) loss curves of denominator D in CNN 1; (c) loss curves of order k in CNN 2

Fig. 9. Output results of CNN. (a) Measured object; (b) wrapping phase obtained by CNN 1; (c) fringe order obtained by CNN 2

Fig. 10. Wrapped phase error diagram of different methods. (a) Wrapped phase error of proposed method; (b) wrapped phase error of WFT method

Fig. 11. Absolute phase error under different methods. (a) Absolute phase obtained by phase shift method combined with gray code method; (b) absolute phase obtained by proposed method; (c) absolute phase error obtained by proposed method compared with traditional method

Fig. 12. Statistical diagram of absolute phase error under different conditions. (a) Fig. 11 (c) image 1; (b) Fig. 11 (c) image 2; (c) Fig. 11 (c) image 3; (d) Fig. 11 (c) image 4

Fig. 13. Phase diagram of open hand process

Fig. 14. Phase diagram of paper tearing process

Set citation alerts for the article
Please enter your email address