Yoshiki Ichioka

Application of Multiple-Image Fourier Transform Spectral Imaging to Measurement of Fast Phenomena

A new interferometric technique that simultaneously acquires all optical data necessary to reconstruct a spectral image is proposed. The optical data is a two-dimensional array 0L two-dimensional interference patterns. The data are acquired by an optical system consisting of a two-dimensional lens array and a Michelson interferometer. The interference patterns are generally distorted by the errors in path differences. These errors can be estimated beL0rehand and corrected in the process of reconstruction of the spectral image. Experimental results obtained with this technique applied to a fast phenomenon are also presented.

Experimental verification of parallel processing on a hybrid optical parallel array logic system

A hybrid digital optical computing system is constructed, a variation of the optical parallel array logic system (OPALS). The OPALS is a general-purpose digital optical computing system based on optical array logic, in which image coding and 2-D correlation are used to achieve parallel logical operations. In the constructed system, 2-D correlation for optical array logic is performed optically with a modified multireflective correlator; the other procedures in optical array logic are achieved by electronics including a TV feedback system. We have verified correct execution of programs written by optical array logic on the system. Although the processing speed of the system is still slow because of the sequential process in electronics, it can be drastically improved by replacing the sequential processing devices with parallel ones.

Optical fractal synthesizer: concept and experimental verification

An application of optical parallel processing in the generation of fractal images is presented. Iterated function systems [M. Barnsley, Fractals Everywhere (Academic, Boston, Mass., 1988), Chap. 3] are the basis of the operation, which can be easily implemented with optical techniques. An optical fractal synthesizer is considered to compute iterated function systems effectively with the advantages of optical processing in data continuity as well as parallelism. As an instance of the optical fractal synthesizer, an experimental system consisting of two optical subsystems for affine transformation and a TV-feedback line is constructed. Several experimental results verify the principle and show the processing capability of the optical fractal synthesizer.

Optical information processing and beyond

This paper first gives a brief review of fundamentals of optical information processing and practical needs. Then recent progress of optical information processing is reviewed, focusing on the advance of novel processing approaches in optical information processing and new processing systems. Next is a description of the recent advance of spatial light modulators (SLMs) for optical information processing. Architectures for optical and optical/electronic hybrid correlation systems and the recently developed correlation filters are reviewed. Correlation experiments implementing real-time and/or feedback processing are demonstrated. Finally, beyond the present state-of-the-art in optical information processing, an effort toward developing an intelligent optical human vision system is briefly described as an example of future research in optical information processing.

P-OPALS: pure optical-parallel array logic system

In this paper, we present an all-optical version of the OPALS, called P-OPALS, as an instance of a pure optical computing system to clarify its potential capabilities. Before discussing the P-OPALS, basic concepts of OAL and the OPALS are explained. On the P-OPALS, utilized techniques and an experimental system are presented along with experimental results. Finally, processing capability of the P-OPALS is estimated in terms of data number and processing capacity. >

Flexible-structured computation based on optical array logic

A flexible processing scheme for various types of structured data is presented. The scheme is based on optical array logic, which is a technique for parallel neighborhood operation using spatial coding and 2-D correlation. In the proposed scheme, virtual processing elements are constructed on an image plane and their functions are emulated with optical array logic. Such a processing system can easily be controlled by programming in optical array logic, so that flexible-structured data processing can be achieved. Several computer simulation results are presented. Referring to the results, requirements for optical implementation are discussed.

Hybrid optical parallel array logic system

A complete parallel version of hybrid optical parallel array logic system (H-OPALS 162) is constructed, which can process 16 X 16 pixels simultaneously at up to 19 kframs/sec. In the system, all pixels are processed and transferred in parallel. Experimental results, including parallel numerical processing and image processing, show capability and flexibility of the system.

Database management using optical array logic

Database management is considered as an application of optical array logic. To generalize the problem, a systematic procedure for massively parallel data processing that consists of pattern expansion, template matching, magnitude comparison, and sorting is presented. With this procedure, a method for implementing basic operations for relational database models is developed. Basic operations are described by parallel programs in optical array logic. Developed programs are evaluated by their performance. To improve the performance, the development of specialized optical functional modules is needed.

Optical implementation of the banyan network using a Sagnac inverter with a patterned mirror

Abstract A new technique of implementing the banyan network based on free-space optics is presented. The banyan network belongs to a class of log2N multistage interconnection networks. A Sagnac inverter with a patterned mirror is used to implement the banyan network. The proposed optical implementation is lossless in both power and resolution, skew-less for every signal, and relatively simple in configuration.

Hybrid high speed pattern matching using a binary incoherent hologram generated by a rotational shearing interferometer.

A hybrid image processing system for high speed pattern matching is presented. The system consists of an incoherent optical system and an electronic digital image processor. The optical system makes use of the matched filtering technique based on incoherent holography. The incoherent holograms of an input object and a reference object are binarized. This binarization of holograms has two major advantages: (1) it sharpens the autocorrelation peak, and (2) it makes real time processing possible. We demonstrate the high potential of this method by presenting computer simulations and experiments.