JP2749944B2 - Optical arithmetic unit - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a high-speed arithmetic device using light, such as a control device.

[Conventional technology]

An example of the prior art will be described with reference to FIG. Binary input image
A1 and B2 are composed of m × n pixels, and pass light corresponding to binary levels 0 and 1 by the method shown in FIG.
The encoded input image 3 is represented by light and shade that does not pass through, and is superimposed to form an encoded input image 3 represented in FIG. 3 (b). This image is placed on the input surface 4 of the projection optical system, and LEDs 6 arranged in a grid on the light source surface 5
Illuminate by. By adjusting the arrangement of the system, the projected images of the coded input image 3 by the four LEDs 6 are projected on the screen 7 so as to be shifted by half a pixel vertically and horizontally. This projected image is observed through a decoding mask 8 having square windows arranged in a square lattice. As a result, the operation result of the two-variable binary logic function is obtained as light light / dark signals in parallel with respect to all pixels constituting the image. By changing the combination of the blinking states of the four LEDs 6 as shown in FIG. 4, 16 types of logical operations can be performed. This method can execute a parallel logical operation based on the parallelism of light.

For more information about this, see "Optical Computer" published by Ohmsha.
It is described on pages 102 to 111.

[Problems to be solved by the invention]

 The problems of the prior art are as follows.

a. To make a negative expression for numerical operation in this way,
It requires encoding at the logical level.

 Therefore, for example, a complement expression is required.

b. Missing part of the image causes an error.

[Means for solving the problem]

The present invention represents numerical values in binary or ternary representation,
A method of allocating one spatial frequency to each digit of the numerical value represented by the expression method, a means for adjusting the amplitude and phase of the light wave for expressing the spatial frequency in the form of a light wave, a plurality of light waves , A means for optically performing a spatial inverse Fourier transform to convert a spatial frequency into a lattice having a cosine or sine wave light amplitude distribution (a sine wave or a cosine wave amplitude lattice), and performing a spatial Fourier transform on the lattice. Means for returning the frequency to the spatial frequency.

[Action]

Numerical values are represented in a binary or ternary representation, and one frequency on a spatial frequency is assigned to each digit. For example, each digit is expressed in such a manner that if one digit is 1, the assigned frequency exists, and if 0, it does not exist. further,
In the ternary representation, if there is a digit of −1, the assigned frequency exists and the phase of the assigned frequency component is adjusted so as to be 180 ° different from the case where it is 1. What expresses the numerical value by the light wave is means for expressing the amplitude and phase of the light wave.

Thus, when the numerical value represented by the amplitude and phase of the light wave is converted by means of optically spatial inverse Fourier transform,
Each digit is a light / dark pattern of a sine wave or cosine wave amplitude grating. In this state, when the two numerical values represented as described above are superimposed, the amplitude grid representing each digit is strengthened or weakened, so that addition and subtraction are performed. Further, when the result is returned to the spatial frequency by means of the spatial Fourier transform, the numerical expression returns to the original numerical value reflecting the result of addition and subtraction.

〔Example〕

One embodiment of the present invention will be described with reference to FIGS. The light wave 9 emitted from the light source 10, which is preferably coherent light, is split by the half mirror 11 and split in two directions. One passes through an ND filter 12, passes through a beam expander 13, the diameter of the light wave is enlarged, and passes through a spatial modulator 14. The spatial modulator 14 is composed of a phase delay material as shown in the principle diagram of FIG.
It is composed of 30, an amplitude control material 31, and a pinhole 32, and modulates an input light wave so as to correspond to 1, 0, -1 shown in the lower half of FIG. With this modulation, binary or ternary
Gives a number represented by a value. This is done for each digit of the given number. The phase delay material 30 can be realized by an electro-optic material or the like, and the amplitude control material 31 can be realized by a filter or the like.

Also, the light wave split into the other side is modulated by obtaining the same process. When these light waves are superimposed on each other by the half mirror 15 and subjected to spatial inverse Fourier transform by the Fourier transform lens 16, each digit of the numerical value becomes an amplitude grating exemplified on the right side in the lower half of FIG. Here, assuming that the focal lengths of the Fourier transform lens 16 and the Fourier transform lens 17 are f16 and f17, respectively, the optical distance between the spatial modulator 14 and the Fourier transform lens 16 is equal to f16, and the spatial modulator 14b and the Fourier transform lens 16 The optical distance via the half mirror 15 is also f16. The interval between the two Fourier transform lenses 16 and 17 is set equal to f16 + f17. Stop 14 and Fourier transform lens 17
Is f17. These have original digits of 1 and-, respectively.
In the case of 1, since the phases of the gratings are different by 180 °, they are canceled each other and their spatial frequency components disappear. Also, if the original digit is 0
In the case of a combination of 1,0 and -1,1 + 1, -1 + -1, each spatial frequency component is stored as it is. Therefore, addition and subtraction are realized by superposition of lattices having a certain spatial frequency.

Further, the result is subjected to a spatial Fourier transform by a Fourier transform lens 17, and when a DC component appearing at the center of the light wave and the other spatial frequency component are superimposed on the light receiving element array 25, a binary or ternary numerical value is obtained. Is obtained,
This is the result of addition or subtraction.

According to the phase distribution control means of the embodiment, 6-bit information can be obtained. That is, a pair of a phase delay material, an amplitude control material, and a pinhole at a target position sandwiching the optical axis,
One bit can be represented by the phase delay material, the amplitude control material, and the pinhole at the optical axis center 26.

〔The invention's effect〕

According to the present invention, the expression of 0, 1, and -1 can be optically realized, so that the operation at the level of the logical expression is unnecessary. Also, since the numerical values are represented as a group of amplitude grids, the original numerical values can be reproduced even if a part of the grids is missing.

[Brief description of the drawings]

FIG. 1 is an overall view of an apparatus according to an embodiment of the present invention, FIG. 2 is a principle diagram of a conventional optical operation device, and FIGS. 3 and 4 are input diagrams of the conventional optical operation device. FIG. 5 is a diagram showing the relationship between the number expression and the spatial inverse Fourier transform by the phase distribution modulator of the present invention.

Claims (1)

(57) [Claims] 1. A light source, comprising: a plurality of phase control means for controlling the phase of a light wave spread from the light source in a plane, wherein spatial cosine or sine wave light provided by the plurality of phase control means is provided. A means for adding a sign to the image of the grating having an amplitude distribution by a spatial frequency, combining the light waves output from the phase control means, a means for performing a spatial (inverse) Fourier transform, and a means for performing a spatial Fourier transform again An optical operation device characterized by the above-mentioned.