16th Australasian Fluid Mechanics Conference Crown Plaza, Gold Coast, Australia 2-7 December 2007 Multi-camera digital holographic PIV: Tomographic DHPIV C. H. Atkinson 1 and J. Soria 1 1 Laboratory for Turbulence Research in Aerospace and Combustion, Department of Mechanical Engineering, Monash University, Victoria, 3800, AUSTRALIA Abstract The wide scale application of digital holographic particle image velocimetry (DHPIV) as a three-component three-dimensional (3C-3D) velocity field measurement tool is current restricted by the limited size and resolution of commercially available CCD arrays, resulting in a elongation of particle is the direction nor- mal to the hologram plane. This elongation can be over an order of magnitude greater than the true particle diameter and posses significant problems for the cross-correlation analysis used in particle image velocimetry (PIV). In this paper we discuss a multi-camera method of tomographic digital holographic parti- cle image velocimetry (Tomo-DHPIV) to reconstruct a 3D in- tensity field without a loss of resolution in the hologram nor- mal direction. Application of this reconstruction technique is provided along with Monte Carlo simulations of the effects of various operating parameters. Introduction The experimental investigation of many flows is limited by an inability to measure their instantaneous three dimensional (3D) structure. This in turn places significant limitations on our un- derstanding of the complex turbulent and unsteady phenomena that are commonly found in geophysical and engineering flows. Digital holographic particle image velocimetry (DHPIV) of- fers arguably the best prospect for a standard three-component three-dimensional (3C-3D) velocity field measurement tool. The advantage of DHPIV comes the inherent three-dimensonal nature of holographic recording (4). Since a hologram records the pattern of interference between light scattered from particles and that of a reference wave, information about both the ampli- tude and phase of scattered light are stored (figure 1a). A holo- gram can then be used to reconstruct the intensity distribution throughout an entire volume (figure 1b). Each hologram there- fore records the entire 3D intensity distribution, as opposed to tomographic technqiues that rely on trying to solve for the 3D intensity distribution based on a multiple 2D images. 3C-3D ve- locity fields can then be determine from pairs of reconstructed holograms using 3D cross-correlation techniques, similar to that of standard planer PIV (2). The use of CCD arrays for holographic recording removes the need for time consuming film processing and enables direct dig- ital holographic reconstruction, without the need for complex optical reconstruction and scanning digitization. This provides a significant step towards mainstream use of DHPIV, yet as with the move from film based PIV to digital PIV, does so at the ex- pense of resolution. Unfortunately this loss of recording resolu- tion is far more serious in DHPIV owing to effect of resolution limited interference fringe spacing on the depth-of-field and ac- curacy normal to the hologram plane (3). In the case of in-line holography of 11 μm diameter spherical particles this accuracy normal to the hologram can be can be on the order of 20 times the particle diameter, resulting in the reconstruction of ellip- soids in the normal direction. This particle elongation can not only obscure other particles, but can also result in cross-talk be- tween planes normal to the viewing direction, with both effects being highly undesirable in cross-correlation PIV analysis. In this paper we discuss a new technique of tomographic dig- ital holographic particle image velocimetry (tomo-DHPIV) (6) where multiple holographic reconstructions or 3D projections from different orientations are combined to remove depth-of- field limitations. By retaining only the region of a particles that fall in the overlapping domain of multiple cameras the depth- of-field bias of each view is removed, resulting in a more ac- curate quasi-spherical particle reconstruction. A discussion of the technique is provided along with a numerical investigation of the optimal operating parameters. Figure 1: Schematic of (a) holographic recording and (b) holo- graphic reconstruction. Digital Holography and Depth-of-Field In digital particle holography the interference pattern created by the light scatter from the particles or the object wave and that of a reference wave is recordered directly onto a CCD array. These waves may either originate from two separate coherent beams or in the case of in-line holography a single beam where the light scattered by the particles forms the object wave and the light that passes through the particle field forms the refer- ence wave. One advantage of digital holography is that once the interference pattern is recorded it is instantly available in a digital form, without the need to develope and then digitise a holographic plate. The other advantage is that using algorithms such as that of Onural and Scott (5) the 3D volume intensity field may be directly calculated from the digital hologram, re- placing the time consumming process of realigning and repro- jecting the reference wave through a developed hologram. 184