Research On The Algorithm Of Three Dimensional Reconstruction Based On Digital Holography

The algorithm of the reconstruction of three dimensional objects based on digitalholography is addressed in this paper. We combine the methods of analyzing the theories andinvestigating the experiments, and research the basic questions and kinds of methods of three,dimensional reconstructions based on digital holography, including the methods of removingthe zero-order terms and a pair of twin-images, the algorithms of digital holographicreconstruction. The digital holograms captured by CCD are mostly Fresnel off-axisholograms, because of the limit of physical process; they contain zero-order terms andtwin-image in the reconstruction of diffraction based on the Fresnel transformation. Thetraditional method limits the field of the vision, even though it can restrict the zero-orderterms and twin-image. In the process of reconstruction based on traditional Fouriertransformation, it adopts the methods of filtration and shift-frequency, the loss of signal isunavoidable and the three-dimensional model can't be reconstructed because of the loss ofthe information of phase, especially there is a frequency spectrum mixture. So, this paperwill focus on the research of the following aspects:First, during the process of capture and reconstruction of holograms, we bring forwardthe method of complex-wave retrieval based on blind signal separation of stochastic signalprocessing. During the research, we find that the process of capturing the interferometrichologram by CCD is stochastic, in this way, the reconstruction of digital holograms can builton the theory of stochastic signal processing. So we research a new method of thereconstruction of digital hologram based on the theory of stochastic signal processing. It canretrieve the phase in the digital holographic reconstruction effectively, separate and removethe zero-order terms and a pair of twin-images instead of restraining and avoiding them, andpreserve the information of the amplitude and phase of complex-wave. The method is nolonger sensitive to the problem of intensity of exposal, definition and wavelength oflight-wave which puzzle the hologram for a long time. The whole process of processing isdone in spatial domain for obtaining the real phase of digital holograms; it reforms themethod of reconstruction of digital hologram, the achievement is valuable in other aspects inthe domain of light information processing.Second, in the field of researching the algorithms of phase unwrapping, we bringforward a new improved algorithm. The digital holography technology uses CCD instead of the recording medium in the traditional optical holography to record the holograms. Becauseof the system, the phase of hologram captured is restricted between (—π,π]. How toretrieve the ambiguous phase period and receive the real unwrapping phase, namely phaseunwrapping processing, which is one of the most important problems in the reconstruction ofdigital holography. There are noises, image shadow and layover, which lead to the phaseinconsistent, named residues, in the real digital hologram. Residues lead to the transfer andaccumulation of the errors by receiving the result which is related to the path, while doingthe simple integral to the measured phase gradient. To obtain the real phase of digitalhologram, we plan to compare the two traditional algorithms of phase unwrapping:branch-cut algorithm of Goldstein and least squares algorithm, and adopt the improvedsimulated annealing algorithm to unwrap the phase of complex-wave, obtain the optimalalgorithm of phase unwrapping, which can effectively solve the problem of digitalholographic reconstruction.