Two-dimensional wavelet transform and application to holographic particle velocimetry

Transform-based phase retrieval techniques from a single off-axis interferogram

Optical phase retrieval (OPR) methods are important because they are used to obtain the transverse phase profile information of a beam. Interference methods are extensively used to convert the phase information into an intensity pattern, which can then be processed further to retrieve the unknown phase. The most widely used interference method involves the interference of the unknown object beam and a known reference beam with an angle between them. There are several algorithms that retrieve the phase information from such a single off-axis interference pattern. For a particular application, the choice of an algorithm for OPR is very important. Therefore, it is necessary to choose between them, depending on the requirements. Three entirely different noniterative, transform-based algorithms, namely the Fourier transform (FT) method, the continuous wavelet transform (CWT) method, and the Hilbert transform (CWT) method, are explained in detail. A quantitative comparison is made using a combination of rms error and standard structural similarity measure. The advantages of using a standard unwrapping algorithm are also validated using the same combination of comparison metrics. We show that the HT method has a better response with object beam with higher spatial frequency content, but with the penalty of affected noise. The FT method and CWT method have better noise immunity, but have the limitation of the spatial frequency range of the object beam. The different constraints, advantages, and some practical limitations of the methods are discussed with the help of a quantitative phase imaging experiment of monodispersed polymethyl methacrylate beads.

Extended wavelet transformation to digital holographic reconstruction: application to the elliptical, astigmatic Gaussian beams

Wavelet analysis provides an efficient tool in numerous signal processing problems and has been implemented in optical processing techniques, such as in-line holography. This paper proposes an improvement of this tool for the case of an elliptical, astigmatic Gaussian (AEG) beam. We show that this mathematical operator allows reconstructing an image of a spherical particle without compression of the reconstructed image, which increases the accuracy of the 3D location of particles and of their size measurement. To validate the performance of this operator we have studied the diffraction pattern produced by a particle illuminated by an AEG beam. This study used mutual intensity propagation, and the particle is defined as a chirped Gaussian sum. The proposed technique was applied and the experimental results are presented.

Two-dimensional wavelet transform for reliability-guided phase unwrapping in optical fringe pattern analysis

This paper theoretically discusses modulus of two-dimensional (2D) wavelet transform (WT) coefficients, calculated by using two frequently used 2D daughter wavelet definitions, in an optical fringe pattern analysis. The discussion shows that neither is good enough to represent the reliability of the phase data. The differences between the two frequently used 2D daughter wavelet definitions in the performance of 2D WT also are discussed. We propose a new 2D daughter wavelet definition for reliability-guided phase unwrapping of optical fringe pattern. The modulus of the advanced 2D WT coefficients, obtained by using a daughter wavelet under this new daughter wavelet definition, includes not only modulation information but also local frequency information of the deformed fringe pattern. Therefore, it can be treated as a good parameter that represents the reliability of the retrieved phase data. Computer simulation and experimentation show the validity of the proposed method.

Wavelet ridge techniques in optical fringe pattern analysis

Wavelet ridge techniques utilizing daughter wavelets under two different kinds of definitions in the optical fringe pattern analysis are theoretically clarified. The clarification reveals that the phase of the optical fringe pattern is equal to that of its wavelet transform coefficients on the ridge using both of the two wavelet definitions. The differences between the two definitions in the performance of wavelet transform algorithms are verified in theory. The strict relations between the instantaneous frequency of the fringe pattern and the scale parameter at the wavelet ridge position are also theoretically clarified for the phase gradient method. A simple method for selecting the scale vector is introduced. Computer simulations and experiments reveal the correctness of the clarification and the validity of the proposed method.

Spatial carrier fringe pattern phase demodulation by use of a two-dimensional real wavelet

A two-dimensional continuous wavelet transform employing a real mother wavelet is applied to phase analysis of spatial carrier fringe patterns. In this method, a Hilbert transform is first performed on a carrier fringe pattern to get an analytic signal. Then a two-dimensional wavelet transform is calculated for the signal that is yielded by the first transform. Finally, the height-demodulated phase information can be gotten from the wavelet transform coefficients at the wavelet ridge position. The performance of the proposed method has been evaluated by using computer-generated and real fringe patterns. The result performed better than that of one-dimensional real wavelet transform algorithms in the area with phase discontinuous points and high phase variation, especially when there is much noise in the fringe patterns. Computer simulations and experiments verified the validity of the proposed method.

Phase reconstruction of digital holography with the peak of the two-dimensional Gabor wavelet transform

We describe a numerical reconstruction technique for digital holography by means of the two-dimensional Gabor wavelet transform (2D-GWT). Applying the 2D-GWT to digital holography, the object wave can be reconstructed by calculating the wavelet coefficients of the hologram at the peak of the 2D-GWT automatically. At the same time the effect of the zero-order diffraction image and the twin image are eliminated without spatial filtering. Comparing the numerical reconstruction of a holographic image by the analysis of the one-dimensional Gabor wavelet transform (1D-GWT) with the 2D-GWT, we show that the 2D-GWT method is superior to the 1D-GWT method, especially when the fringes of the hologram are not just along the y direction. The theory and the results of a simulation and experiments are shown.

Application of the fractional Fourier transformation to digital holography recorded by an elliptical, astigmatic Gaussian beam

The authors have studied the diffraction pattern produced by a particle field illuminated by an elliptic and astigmatic Gaussian beam. They demonstrate that the bidimensional fractional Fourier transformation is a mathematically suitable tool to analyse the diffraction pattern generated not only by a collimated plane wave [J. Opt. Soc. Am A 19, 1537 (2002)], but also by an elliptic and astigmatic Gaussian beam when two different fractional orders are considered. Simulations and experimental results are presented.

Hologram reconstruction by use of optical wavelet transform

High-speed in-line holography is used to visualize the trajectories of glass fibers being drawn out in a turbulent flame. To improve the signal-to-noise ratio, the images are not observed by a conventional reconstruction setup, but the holographic plate is placed directly on the input plane of a wavelet-transform optical system. This processing system is based on a VanderLugt correlator with inclusion of an electrically addressed spatial light modulator. The shape of the matched filters is deduced by successive rotation and dilatation operations of wavelet functions in the Fourier domain. We estimate the three-dimensional location of a fiber element and its orientation by searching for the daughter wavelet that yields the maximum intensity on the output plane of the correlator, which also contains the reconstructed image. The results are compared with those obtained by conventional optical reconstruction. The signal-to-noise ratios of the images observed on the output plane are improved. Moreover, it is shown that the axis coordinate accuracy is improved to Dz = +/-50 microm, instead of +/-0.5 mm for holographic reconstruction.