There has been much recent interest in the application of optical tomography to the study of transport phenomena and chemical reactions in transparent fluid flows. An example is the use of multidirectional holographic interferometry and computed tomography for the study of crystal growth from solution under microgravity conditions. A critical part of any such measurement system is the computed tomography program used to convert the measured interferometric data to refractive index distributions in the object under study. Several of the most promising CT algorithms for this application are presented and compared here. Because of the practical difficulty of making multidirectional interferometric measurements, these measurements generally provide only limited amounts of data. Recent studies have indicated that of the several classes of reconstruction algorithms applicable in the limited-data situation, those based on the Multiplicative Algebraic Reconstruction Technique (MART) are the fastest, most flexible, and most accurate. Several MART-type algorithms have been proposed in the literature. In this paper we compare the performance of state-of-the-art implementations of four such algorithms under conditions of interest to those reconstructing multidirectional interferometric data. The algorithms are tested using numerically-generated data from two phantom objects, with two levels of added noise and with two different imaging geometries. A reconstruction of real data from a multidirectional holographic interferometer using the best of the algorithms is shown.

Date Published: 1 January 1992
PDF: 12 pages
Proc. SPIE 1553, Laser Interferometry IV: Computer-Aided Interferometry, (1 January 1992); doi: 10.1117/12.135322

Published in SPIE Proceedings Vol. 1553:
Laser Interferometry IV: Computer-Aided Interferometry
Ryszard J. Pryputniewicz, Editor(s)