• Acta Optica Sinica
  • Vol. 38, Issue 7, 0712002 (2018)
Cheng Zhang and Wenjing Chen*
Author Affiliations
  • College of Electronics and Information Engineering, Sichuan University, Chengdu, Sichuan 610065, China
  • show less
    DOI: 10.3788/AOS201838.0712002 Cite this Article Set citation alerts
    Cheng Zhang, Wenjing Chen. Method for Improving Measurement Accuracy of Wavelet Transform Profilometry[J]. Acta Optica Sinica, 2018, 38(7): 0712002 Copy Citation Text show less
    Optical geometry of measurement system
    Fig. 1. Optical geometry of measurement system
    (a) Daughter wavelets with different scales; (b) Fourier spectra of daughter wavelets; (c) mother wavelets with different fb values; (d) Fourier spectra of mother wavelets
    Fig. 2. (a) Daughter wavelets with different scales; (b) Fourier spectra of daughter wavelets; (c) mother wavelets with different fb values; (d) Fourier spectra of mother wavelets
    (a) Deformed fringe pattern without noise; (b) wavelet ridges with different fb values; (c) unwrapped phase of Fig. 3(b); (d) wavelet ridges when fb is equal to 0.2; (e) unwrapped phase of Fig. 3(d)
    Fig. 3. (a) Deformed fringe pattern without noise; (b) wavelet ridges with different fb values; (c) unwrapped phase of Fig. 3(b); (d) wavelet ridges when fb is equal to 0.2; (e) unwrapped phase of Fig. 3(d)
    (a) Deformed fringe pattern considering random noise with mean square deviation of 0.20; (b) wavelet ridges with different fb values; (c) unwrapped phase of Fig. 4(b); (d) wavelet ridges introducing improved wavelet processing method; (e) unwrapped phase of improved wavelet processing method
    Fig. 4. (a) Deformed fringe pattern considering random noise with mean square deviation of 0.20; (b) wavelet ridges with different fb values; (c) unwrapped phase of Fig. 4(b); (d) wavelet ridges introducing improved wavelet processing method; (e) unwrapped phase of improved wavelet processing method
    (a) Simulated object; (b) deformed fringe pattern
    Fig. 5. (a) Simulated object; (b) deformed fringe pattern
    Error distribution without noise. (a) fb=0.5; (b) fb=1.0; (c) fb=1.5; (d) fb=2.0
    Fig. 6. Error distribution without noise. (a) fb=0.5; (b) fb=1.0; (c) fb=1.5; (d) fb=2.0
    Error considering random noise influence with mean square deviation of 0.20.(a) fb=0.5; (b) fb=1.0; (c) fb=1.5; (d) fb=2.0
    Fig. 7. Error considering random noise influence with mean square deviation of 0.20.(a) fb=0.5; (b) fb=1.0; (c) fb=1.5; (d) fb=2.0
    Results and error distribution considering random noise with mean square deviation of 0.20. (a) Reconstruction results by improved wavelet processing method; (b) reconstruction error distribution by improved wavelet processing method; (c) error distribution by method proposed by Liu et al.[8]
    Fig. 8. Results and error distribution considering random noise with mean square deviation of 0.20. (a) Reconstruction results by improved wavelet processing method; (b) reconstruction error distribution by improved wavelet processing method; (c) error distribution by method proposed by Liu et al.[8]
    (a) Deformed fringe with low noise; (b) fringe distribution of the 381th row in Fig. 9(a)
    Fig. 9. (a) Deformed fringe with low noise; (b) fringe distribution of the 381th row in Fig. 9(a)
    (a) Unwrapped phase by employing improved wavelet processing method; (b) unwrapped phase of the 381th row of deformed fringe pattern
    Fig. 10. (a) Unwrapped phase by employing improved wavelet processing method; (b) unwrapped phase of the 381th row of deformed fringe pattern
    (a) Deformed fringe pattern with high noise level; (b) fringe distribution of the 286th row in Fig. 11(a)
    Fig. 11. (a) Deformed fringe pattern with high noise level; (b) fringe distribution of the 286th row in Fig. 11(a)
    (a) Wrapped phase when fb=1.0; (b) wrapped phase when fb=2.0; (c) unwrapped phase when fb=1.0; (d) unwrapped phase when fb=2.0; (e) wrapped phase extracted by method proposed by Liu et al.[8]; (f) wrapped phase extracted by improved wavelet processing method; (g) unwrapped phase extracted by method proposed by Liu et al[8]; (h) unwrapped phase extracted by improved wavelet processing method
    Fig. 12. (a) Wrapped phase when fb=1.0; (b) wrapped phase when fb=2.0; (c) unwrapped phase when fb=1.0; (d) unwrapped phase when fb=2.0; (e) wrapped phase extracted by method proposed by Liu et al.[8]; (f) wrapped phase extracted by improved wavelet processing method; (g) unwrapped phase extracted by method proposed by Liu et al[8]; (h) unwrapped phase extracted by improved wavelet processing method
    Mean squaredeviation of noiseStandard deviation for improvedwavelet processing method /mmStandard deviation for complex Morlet wavelet method /mm
    fb=0.5fb=1.0fb=1.5fb=2.0
    00.07000.06990.10210.12970.1520
    0.050.07760.07660.10560.13200.1537
    0.100.09450.09350.11500.13840.1587
    0.150.11430.11540.12760.14650.1645
    0.200.13920.26450.26650.15970.1753
    Table 1. Standard deviation of reconstruction errors for complex Morlet wavelet method with different fb values and improved wavelet processing method at different noise levels
    Cheng Zhang, Wenjing Chen. Method for Improving Measurement Accuracy of Wavelet Transform Profilometry[J]. Acta Optica Sinica, 2018, 38(7): 0712002
    Download Citation