KR101270768B1 - Hologram playback device and method - Google Patents

Abstract

The present invention relates to a hologram reproducing apparatus and a reproducing method, and more particularly, to a hologram reproducing apparatus and a reproducing method having a high processing speed using a distributed computer.
To this end, the hologram reproducing apparatus of the present invention divides a recording medium on which a fringe pattern for reproducing a hologram is recorded, a fringe pattern recorded on the recording medium, into segments having a predetermined size, and the fringe patterns recorded on the divided segments. And a fast Fourier transform unit for performing fast Fourier transform, and an adder for adding fringe patterns of a plurality of segments on which the fast Fourier transform has been performed.

Description

Hologram display device and method {Hologram display, Method for hologram display}

The present invention relates to a hologram reproducing apparatus and a reproducing method, and more particularly, to a hologram reproducing apparatus and a reproducing method having a high processing speed using a distributed computer.

Three-dimensional image realization technology that allows you to feel three-dimensional depth and three-dimensional feeling from a planar image is widely applied to the aerospace, arts, and automotive industries, as well as the home appliance and telecommunications industry, as well as the direct related fields such as displays. The technical ripple effect is expected to be more than the current HDTV.

That is, with the development of the digital industry and the image industry, high resolution HDTVs have been developed to satisfy human visual effects, and stereoscopic display devices for displaying 3D stereoscopic images have been developed in various ways.

The most important factor for human beings to feel depth and stereoscopic effect is the binocular disparity due to the interval between the two eyes, but there is also a deep relationship with the psychological and memory factors. Therefore, A three dimensional representation method (holographic type), and a stereoscopic type (stereoscopic type) based on whether three-dimensional image information of the three-dimensional image can be provided.

The volume expression method is a way to feel the perspective of the depth direction due to psychological factors and inhalation effects. The viewing angle is calculated by a 3D computer graphic or an observer who displays perspective, superposition, shadow, contrast, and movement by calculation. It is applied to so-called IMAX movies, which provide a large large screen and cause optical illusions to be sucked into the space.

The three-dimensional representation method known as the most complete stereoscopic image implementation technology can be represented by laser light reproduction holography to white light reproduction holography.

In addition, the stereoscopic expression method is to sense a three-dimensional feeling using physiological factors of both eyes. Specifically, when a plane-related image of parallax information is seen in the left and right of human beings that are about 65 mm apart, the brain merges them. The ability to generate spatial information before and after the display surface and to sense a three-dimensional effect, that is, using stereography. This stereoscopic effect expression system is called a multi-view display system. Depending on the position of actual stereoscopic effect generation, a spectacle system using special glasses on the observer side or a parallax barrier, a lenticular, or an integral system on the display surface side, And a non-eyeglass system using a lens array such as a lens array.

The integrated image method, which is one of the volumetric representation methods, reproduces the optical characteristics that are the same as the distribution and luminance of the light emitted from the actual three-dimensional object, thereby allowing the virtual three-dimensional stereoscopic image to be recognized even if there is no actual three-dimensional object.

These methods can be divided into the presence or absence of wearing special glasses when observing three-dimensional stereoscopic image, and when viewing the three-dimensional image, without wearing special glasses, seeing the natural three-dimensional image, do not feel fatigue even if observed for a long time The main goal is to develop a 3D stereoscopic image display method.

As described above, the holographic display is known as one of the ideal stereoscopic image display methods.

Hologram (HOLOGRAM) is a combination of HOLOS, which means Greek, and GRAM, which means message, and is obtained by flattening information of visible image (stereosity). In other words, two-dimensional information is two-dimensionalized and reproduced again from the two-dimensional information into a three-dimensional image. Holograms use the coherence of light reflected from an object, and the reflected wavefront of light, including phase and amplitude, appears in three dimensions, which is recorded in a two-dimensional optical refraction medium. Say it as it is. Holography is a photography that records the interference wave generated by the interference of the object beam and the reference beam reflected from the three-dimensional object. A hologram film using a high particle is used to generate the hologram, and the interference pattern is recorded using monochromatic light.

Fig. 1 shows a fringe pattern recorded on a hologram recording device. As described above, the hologram recording apparatus records the fringe pattern for reproducing the hologram on the recording medium using the object beam and the reference beam.

The hologram reproducing apparatus reproduces the hologram using the fringe pattern recorded on the recording medium. In general, the fringe pattern is large and dynamic, with a size of 100k ㅧ 100K or 1M ㅧ 1M. Therefore, the hologram reproducing apparatus requires a system having a high speed data processing capability in order to reproduce a large fringe pattern.

An object of the present invention is to propose a hologram reproducing apparatus and a reproducing method having a high speed data processing capability.

Another object of the present invention is to propose a method for reproducing holograms having the same size regardless of the size of the fringe pattern.

To this end, the hologram reproducing apparatus of the present invention divides a recording medium on which a fringe pattern for reproducing a hologram is recorded, a fringe pattern recorded on the recording medium, into segments having a predetermined size, and the fringe patterns recorded on the divided segments. And a fast Fourier transform unit for performing fast Fourier transform, and an adder for adding fringe patterns of a plurality of segments on which the fast Fourier transform has been performed.

To this end, the hologram reproducing apparatus of the present invention divides a fringe pattern recorded on a recording medium on which a fringe pattern for reproducing a hologram is recorded into segments having a predetermined size, and performs fast Fourier transform on the fringe pattern recorded on the divided segments. Parallel control module for performing a; and a CPU core for adding the fringe pattern of the plurality of segments subjected to the fast Fourier transform.

To this end, the hologram reproducing method of the present invention divides a fringe pattern recorded on a recording medium on which a fringe pattern for reproducing the hologram is recorded into segments having a predetermined size, and a high-speed fringe pattern recorded on the divided segments. Performing a Fourier transform, adding a fringe pattern of the plurality of segments on which the fast Fourier transform is performed.

The hologram reproducing apparatus and the reproducing method according to the present invention have a high speed data processing capability and can reproduce holograms having the same size regardless of the size of the fringe pattern.

1 shows a general fringe pattern recorded on a recording medium,
2 shows a configuration for reproducing holograms using a fringe pattern recorded on a conventional recording medium,
3 illustrates a configuration for reproducing a hologram using a fringe pattern recorded on a recording medium according to an embodiment of the present invention.
4 illustrates an example in which a fringe pattern is divided into a plurality of segments according to an embodiment of the present invention.
5 illustrates a distributed computer for processing a fringe pattern recorded on a recording medium according to an embodiment of the present invention.

The foregoing and further aspects of the present invention will become more apparent through the preferred embodiments described with reference to the accompanying drawings. Hereinafter will be described in detail to enable those skilled in the art to easily understand and reproduce through this embodiment of the present invention.

Today, the increased hardware complexity associated with higher clock frequency processors is beginning to be recognized. In addition, the clock frequency cannot be increased indefinitely, and other methods are required. As a result, multiprocessor and multithreading technologies have emerged as a way of improving the overall efficiency through thread level parallelism (TLP) as well as improving processor efficiency.

The recent shift to multiprocessing has spread from desktop to embedded design, requiring a shift in thinking about many software. For many years, embedded designers have been able to embed more processors into their designs to achieve better computational power with limited power.

Both multiprocessor and multithreading improve processor performance overall, which reduces the processing time of the application by using parallel software threads. However, these two technologies take different approaches in hardware to achieve these goals and represent different levels of achievement for specific software code examples.

The basic concept of multithreading is to increase overall processor performance by taking advantage of the inefficient cycles that typically occur in uniprocessor designs due to access delays when high frequency processors are combined with slow memory.

By fitting threads into spare cycles, core efficiency is improved. By default, multithreading is classified as a uniprocessor that supports additional hardware threads by overlapping only minimal processor logic.

Typically this is a set of registers from the programmer, and the CPU's supervisor state is sufficient so that today's operating systems can see this hardware thread as a virtual processor. Sharing this remaining processor logic is an important issue that increases the complexity of the software.

In general, a CPU is designed to process a sequence of instructions that execute in sequence. The advent of multithreading capabilities and multiple cores allows a significant degree of parallelism in the CPU, but there is a performance difference between the CPU and the GPU.

GPUs are parallel processors designed to handle multiple tasks at the same time. Each GPU core is simpler and less powerful than a CPU core, but the GPU contains dozens or hundreds of cores.

The present invention proposes a method of processing a fringe pattern using a GPU. As described above, in order to reproduce the hologram using the fringe pattern, a fast Fourier transform (FFT) is performed.

2 is a diagram illustrating a process of reproducing a hologram using a conventional fringe pattern. According to Fig. 2, the hologram reproducing apparatus performs fast Fourier transform on all pixels of all regions of the recording medium on which a fringe pattern is recorded. However, in order to reproduce the hologram by performing fast Fourier transform on all the pixels of all regions of the recording medium, a system having a high speed data processing capability is required. As described above, if the fringe pattern has a high resolution, a system having a high speed data processing capability is further required to reproduce the hologram.

However, the resolution required for a person to identify the hologram does not require the high resolution that the fringe pattern has. That is, the hologram can be identified only by a resolution lower than the high resolution of the fringe pattern.

Accordingly, the present invention divides the fringe pattern into segments having a constant size, and performs fast Fourier transform on each segment. It can be seen that the fringe pattern for reproducing the hologram by Equation 1 is reproduced by the fast Fourier transform, and the fast Fourier transform has linearity and phase mobility.

[Equation 1]

h: hologram

r: reference

w: chirp function

T: Shift interval

3 is a diagram illustrating a process of reproducing a hologram using a fringe pattern according to an embodiment of the present invention. That is, a hologram reproducing apparatus for implementing Equation 1 is shown.

According to Fig. 3, the hologram reproducing apparatus includes a plurality of FFT processing units, a plurality of phase shifters and an adder. As described above, the fringe pattern is divided into segments having a constant size, and fast Fourier transform is performed on each segment.

4 illustrates an example of dividing a fringe pattern according to an exemplary embodiment into segments having a predetermined size. The size of the segment should ensure the minimum size at which a human can identify the hologram. In other words, the fringe pattern is divided to ensure a minimum size that can be identified by humans.

In this manner, holograms having the same resolution can be reproduced regardless of the resolution of the fringe pattern. That is, by adjusting the number of segments to be divided according to the resolution of the fringe pattern, holograms having the same resolution can be reproduced.

As described above, the present invention does not perform fast Fourier transform in one CPU, but performs fast Fourier transform using a GPU, that is, a distributed computer.

5 illustrates a system for processing a fringe pattern to reproduce a plurality of holograms according to an embodiment of the present invention.

Referring to FIG. 5, the fringe pattern processing system includes a CPU core 500 and a plurality of parallel control modules 520 constituting the GPU core 510. Of course, it is obvious that other configurations can be included in the fringe pattern processing system.

The CPU core 500 analyzes the input fringe pattern. In other words, the CPU core 500 analyzes the number of segments to be processed. The CPU core 500 determines the number of parallel control modules to be used dynamically to efficiently process the input fringe pattern. The CPU core 500 determines the number of parallel control modules to be processed according to the number of segments constituting the input fringe pattern. The CPU core 500 stores in the log storage the number of parallel control modules determined to process the fringe pattern and the segments to be processed by each parallel control module. The CPU core 500 provides information about the segments constituting the fringe pattern to be processed by each parallel control module.

The CPU core 500 collects data processed by each parallel control module based on information stored in a log storage. If there is a delayed GPU core task while checking the maximum timeout time during the gathering stage, check the corresponding parallel control module identifier to increase the priority of the task so as not to exceed the maximum timeout.

That is, when processing time is used in a specific parallel control module among a plurality of parallel control modules, the CPU core 500 may not perform a process of collecting processed data. Therefore, when the data processing speed is delayed in a specific parallel control module, the CPU core 500 increases the priority of the data processing task of the corresponding parallel control module so that the data can be processed quickly.

In addition, when a data processing time is delayed in a specific parallel control module, the CPU core 500 may request that a parallel control module that does not perform a current data processing task process some or all of the corresponding data. In this case, the CPU core stores new information about it in the load store. By doing so, the CUP core 500 can process the input fringe pattern as quickly as possible.

The parallel control module 520 performs a fast Fourier transform and phase shift process for each requested segment. The parallel control module 520 transfers data obtained by performing a fast Fourier transform and a phase shift process to the CPU core 500. In addition, the parallel control module 520 may provide the CPU core 500 with information about this when the currently requested data processing task is overloaded. In this manner, the CPU core 500 can quickly allocate and request the data processing task to another parallel control module.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the scope of the present invention .

500: CPU core 510: GPU core
520: parallel control module

Claims (9)

delete delete The fringe pattern recorded on the recording medium on which the fringe pattern for reproducing the hologram is recorded is divided into segments having a predetermined size, and at least one segment of the divided segments is provided for fast Fourier transformation on the recorded fringe pattern. At least two parallel control modules including a fast Fourier transform unit performing a phase shifter and a phase shifter performing phase shifting;
A CPU core that adds a fringe pattern of segments output from each parallel control module that has performed data processing including fast fast Fourier transform and phase shift,
The CPU core analyzes the fringe pattern to determine the number of parallel control modules to provide segmented segments,
The time required to perform data processing on the fringe pattern recorded in the segment provided by the specific parallel control module is provided to add the fringe pattern of the output segment from the point of time when the CPU core provides the segment to the parallel control module. If the maximum timeout, which is a period up to a time point, is exceeded, the CPU core provides at least a portion of a segment to be provided to the specific parallel control module to a parallel control module that has not currently performed data processing, or the parallel control module A hologram reproducing apparatus, characterized by providing a CPU core with or without overload of data processing or by setting a priority of data processing of the specific parallel control module relatively high.
delete The method of claim 3, wherein the CPU core,
And storing information about the segment to be processed by the parallel control module and each parallel control module in a log storage.
The method of claim 5, wherein the log storage,
And an identifier of the parallel control module and information on a segment processed by each parallel control module.
4. The hologram reproducing apparatus according to claim 3, wherein the number of the divided segments varies depending on the resolution of the fringe pattern recorded on the recording medium.
Dividing the fringe pattern recorded on the recording medium on which the fringe pattern for reproducing the hologram on the CPU core is recorded into segments having a set size;
Performing at least two fast Fourier transforms and a phase shift in at least two fast Fourier transform units for the recorded fringe patterns by receiving at least one of the segments divided by the parallel control module;
Receiving and adding a fringe pattern of a plurality of segments in which the CPU core performs data processing including the fast Fourier transform and a phase shift,
The CPU core analyzes the fringe pattern to determine the number of parallel control modules to provide segmented segments,
The time required to perform data processing on the fringe pattern recorded in the segment provided by the specific parallel control module is provided to add the fringe pattern of the output segment from the point of time when the CPU core provides the segment to the parallel control module. If the maximum timeout, which is a period up to a time point, is exceeded, the CPU core provides at least a portion of a segment to be provided to the specific parallel control module to a parallel control module that has not currently performed data processing, or the parallel control module The CPU core provides an overload for data processing or sets a priority of data processing of the specific parallel control module to a relatively high level. delete

Images