Author Affiliations
School of Physical Science and Technology, Southwest Jiaotong University, Chengdu 610031, Sichuan , Chinashow less
Fig. 1. Diagram of measuring optical path
Fig. 2. Simulated 3D drawing of rail surface
Fig. 3. Standard sinusoidal fringe image. (a) Fringe image; (b) intensity distribution
Fig. 4. Sinusoidal fringe image with nonlinearity. (a) Fringe image; (b) intensity distribution
Fig. 5. Sinusoidal fringe after twice Hilbert transforms. (a) Intensity distribution; (b) filtered part
Fig. 6. Stoilov phase shift algorithm. (a) Reconstruction result; (b) reconstruction error
Fig. 7. Two-step phase-shift method without compensating phase error. (a) Reconstruction result; (b) reconstruction error
Fig. 8. Two-step phase-shift method with compensating phase error. (a) Reconstruction result; (b) reconstruction error
Fig. 9. Comparison of cross section of defect. (a) Stoilov phase shift algorithm; (b) two-step phase-shift method without compensation; (c) two-step phase-shift method with compensation
Fig. 10. Rail deformation grating fringes required by Stoilov algorithm
Fig. 11. Grating fringes obtained by the proposed method after twice Hilbert transforms
Fig. 12. Wrapped phase maps. (a) Wrapped phase map of Stoilov method; (b) wrapped phase map of proposed method
Fig. 13. Experimental results of three-dimensional profile restoration of rail side. (a) Reconstruction result of Stoilov method; (b) reconstruction result of proposed method
Phase shift algorithm | Maximum absolute error | Root mean squared error | Mean absolute error |
---|
Stoilov phase shift algorithm | 0.8218 | 0.1866 | 0.1489 | Two-step phase-shift algorithm without compensated phase error | 0.7345 | 0.0809 | 0.0616 | Two-step phase-shift algorithm with compensated phase error | 0.6678 | 0.0457 | 0.0324 |
|
Table 1. Error comparison of reconstruction results