JPS62180694A - Three-dimensional television system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method and apparatus for producing television images that are perceived by viewers as three-dimensional, and is used in the television industry.

[Prior art]

Conventional stereoscopic television systems usually require special viewing aids 5, such as polarized or tinted glasses, or special receiving equipment. A representative example of the latter method is disclosed in U.S. Patent Application No. 3.1157.3611.
This is a method of creating 3D television using alternating images from a 3D television camera, as described by J. D. Arrilo. Kaliguchi's method requires a special picture tube for color. Testing has shown that the alternation method described by Kaliguchi (approximately 60 alternations per second, synchronized to television field speed, 30 for each video) has a lower depth than the method using lower alternation speeds. It may reduce the illusion of

U.S. Patent Application No. 66.722 filed on August 25, 1970 and No. 3 filed on November 20, 1972 by the inventor of the present application.
No. 08.209 describes a method for creating the desired five-dimensional effect while remaining compatible with existing broadcast standards and existing home television receivers. However, when two stereoscopic cameras are focused on the foreground φ body according to this method, the background objects may flicker and appear to fly. Similarly,
When the camera is focused on a background object, foreground objects may appear to flicker and fly. The main contributors to this flicker and jump effect are the limitations of binocular fusion in the visual perceptual process combined with the low alternating speed flicker effect. The limits of binocular fusion were first discovered in 1856 by Panum.
) is a visual perception scientist studied by Walter Walter
It is known as a limited case of the Banham phenomenon described in "Perceptual Transmission" edited by A. Rosenprits, published by John Willi and Sons (1961), and the publication described binocular fusion in the 32nd child. The physiological basis for perception is discussed as a result of the alternation of stereoscopic images.

(11I) In an effort to reduce the deleterious effects of this limitation of binocular fusion while preserving the three-dimensional effect, the inventors of the present patent application no.
The invention described in No. 1.006.291 was made. However, while the method of the present invention was designed to minimize the aforementioned effects, it did not completely correct the problem because it did not modify or eliminate its cause. Kaliguchi's method does not have flicker problems since the alternating speed is well above the critical flicker frequencies described in the visual perception literature. However, the limit of the binocular interfusion region appears as a double image.

Basic dimensional illusion methodology □ For humans to visually perceive the 5th dimension, we need stereo images seen from slightly different angles (one image corresponds to each eye), which corresponds in part to the distance between the two eyes. ) is required. This causes each eye to see a slightly different image. Therefore, most five-dimensional movies and five-dimensional television systems require the viewer to use some kind of device, such as polarized glasses, to make the viewer view the left camera image with the left eye and the right camera image with the right eye. Requires the use of color filter glasses or a mechanical shutter viewer.

However, a slightly less convenient method is to simply expose the two stereo images to the viewer's eyes alternately.

A pair of stereoscopic magic lantern images are created by presenting the viewer's eyes, one image partner at a time, alternating first one and then the other several times per second. The normal visual perceptual process interprets images as a single dimensional image insofar as the images are presented alternately such that binocular fusion can occur through its adaptive and integrative abilities and the psychology of the visual process.

The visual perceptual system does not assist in determining which eye sees which image. As long as there is a difference between the two images, and as long as the two images are within a certain range of corresponding binocular positions.

Correct depth relationships are perceived. The visual cortex apparently measures the difference between two images in order to measure the depth of an object. The depth illusion device must have some means of presenting images so that the visual cortex can make this measurement.

There is a simplification in the first paragraph of this section that requires clarification to the fifth order because this clarification is relevant to the present invention. The idea that three-dimensional movies or television requires stereo cameras to be spaced horizontally apart corresponding to the horizontal spacing of the eyes is not entirely correct.

Experiments have shown that it is possible to separate the two cameras vertically (or diagonally) and still achieve the illusion of depth. This result is corroborated by tilting one's head to the side while watching a five-dimensional movie; the stereoscopic illusion of depth still exists even when the head is tilted a full 90 degrees. This perception, which corresponds to a difference in vertical direction, was also developed by O. J. Bradeicz.
ck) and A.C. Sleigh (eds.), “Physical and Biological Processing of Images”; J. Mayhew, “Stereoscopic ” and John Pettigrew
1 Neurophysiology of Binocular Vision”, Scientific American.

This is supported by visual perception literature such as August 1972.

No. 30g, 209 (19
The stereo images are alternated at a rate near or below the critical flicker frequency. This means that some retinal receptors can discharge to the visual cortex under the influence of one stereo image, and other retinal receptors can discharge to the cortex in a time-multiplexed manner under the influence of the other image. It looks like this. Since the divine charge lasts for a short time, the different charges from different stereo images can interact in the binocular fusion process, resulting in the perception of depth.

Our alternating speed is variable, but generally should be on the order of 25 to 25 degrees per second for each image. two partners (one of each)
The relative exposure times of the images can be changed from each equal exposure time to the subthreshold exposure technique of the inventor's US patent application Ser. No. 11,006,291 for significantly unequal exposure times.

However, when two cameras are focused on the foreground object,
Background objects may appear to bounce18). Similarly, when two cameras are focused on a background object, the foreground object may appear to bounce. Problems that appear to be caused solely by flicker phenomena are caused by the combination of flicker phenomena and the limitations of binocular fusion first described by Panum. It should be emphasized here that a binocular image that is focused does not flicker or splatter, while an unfocused binocular image does flicker or splatter.

The difference in the characteristics of Kaliguchi's method and Imsando's method suggests a very fundamental difference in the way the two methods enter into and act on the perceptual process. Since the strain rate of the pothole is well above the critical flicker frequency, there is no flicker problem.

However, this suggests that all retinal receptors are affected by both stereo images and discharge to the visual cortex.

Kaliguchi's monochromatic method also uses spatial separation by placing the two stereo images on alternating interlacing lines, thus ensuring that the two images exist on a spatially separated basis rather than on a true time-multiplexed basis. The resolution, focal length and electron beam sharpness of color picture tubes were not adequate at the time of Kaliguchi's invention to provide the spatial separation needed for color television applications. Therefore, Kaliguchi can use two contrasting colors mixed with a special color picture tube to make the two different stereo images separate in the perceptual cortex so that the binocular image can occur. I discovered that it was necessary.

The method and apparatus described herein for solving the flicker problem of Imsand's method can also be used to solve the double image problem of Kaliguchi's method.

To better explain the invention, aspects of the depth perception instrumentation associated with the invention will now be discussed. The role of the monocular for depth perception is not relevant to the present invention and is therefore not included.

Binocular Elements of Depth Perception Two important cues for visual depth perception are the binocular cues of vergence and stereopsis.

When an intense object is located at a great distance, the lines of gaze from the observer's separate eyes to the object are nearly parallel. When the object is close, the observer's eyes turn toward the object, and the lines of gaze persue each other at an angle that is obvious. When a person fixes their eyes on a finger at arm's length and moves the finger toward their nose while looking with their eyes, the two eyes intertwine. This “more” of the eyes
This can be detected by the perceptual control system that controls the position of the eyeballs, creating the impression that the line is wider or smaller depending on the size of the convergence angle of the eyes. However, vision scientists generally agree that convergence plays a relatively minor role in depth perception. The important result of the ring of convergence is that
Presumably it produces two right and left eye images of the object gazed on, each in the central retinal area of the b eye (very close together).
It is also helpful to locate the corresponding retinal points.

When the eye fixates on a point, geometric analysis can show the theoretical trajectory of all points whose images fall precisely on the corresponding point on the retina. When the eye fixates on a point at the same height as the eye, the locus of the point in the horizontal plane lies on the circumference of a circle that passes through both eyes on one side and through the convergence point on the other side. . When this circle is rotated about an axis passing through the eyes, the resulting donut-shaped solid surface defines this trajectory of points in three spaces. The resulting trajectory 1l-j:5 of each point is clearly not at a constant distance from the eye. However, since binocular fusion only occurs in a small area of the central retina, for practical purposes, the trajectory of each point of accurate retinal correspondence is considered to be equidistant and near the point where the eyes fixate. It will be done.

Stereoscopic Vision When a person looks at an object, the retinal image in the right eye is different from the retinal image in the left eye. This difference is the result of the two eyes viewing objects from two slightly different positions. Experiments have shown that the human visual perceptual system is extremely sensitive to differences between two retinal images. The visual perceptual system uses the amount of difference as a measure of the depth of the object being viewed, and recognizes that the object is getting closer as the difference increases. If there is no difference, it is recognized as a distant background object. Research on visual perception has shown that stereopsis plays a much more important role in depth perception than vergence.

Binocular fusion When an object is viewed with both eyes, the two retinal images are different, but only a single image is usually perceived. This phenomenological cancellation that occurs within the visual cortex of the sensory system is known as binocular fusion.

Limitations of Binocular Fusion When both eyes are focused on an object several feet apart, two slightly different images are seen by both eyes.

Only one binocular fusion is perceived. If the second object is in close proximity to the first object, that too.

Perceived as binocular fusion. When the second object is moved further into the background until the eye is fixed on the first object, a simple geometric projection analysis (see Figures 1A and 18) reveals the background object. It can be seen that the difference in retinal corresponding points in the retinal image becomes sharper. If the difference becomes sufficiently large, the perceptual system will no longer fuse the object binocularly, and a double image will result. When this happens, the limit of binocular fusion has been reached. This is from Murch, Gerald M., Audiovisual Perception, Bobbs Merrill Company.
Merrill] Company), published in 1973,
This is discussed in Chapter 5 as the ``Panam limit case''.

When you straighten your eyes and fixate on the background object, the background object becomes binocularly fused again, and the foreground object becomes a double image (
(See Figures 1A and 18).

Accommodative observers are usually unaware of double images, even though they are present in most complex scenes.

This is because when you focus your eyes on a foreground object, your eyes focus on that object, and the double image of the background object is out of focus and automatically de-emphasized due to the accommodative nature of the visual perceptual process. be. One book points out that accommodation is the focusing of the lens of the eye. However, apemphasis that can occur within the visual cortex
sis) also makes the observer unaware of the out-of-focus double image.

The interaction of vergence and binocular fusion in binocular depth perception is recognized as a complex process in the visual perception literature [Julez, Be1a, Fundamentals of Monocular Perception], University of Chicago Press ( (1971)].However, the following simplifications are consistent with the visual perception literature and are directly relevant to the present invention.

1. The visual perceptual system controls ocular convergence in a way that seeks to maximize the correlation (or retinal correspondence) between left and right eye images of the object of interest.

2. The resulting two different images are processed by the visual cortex and the two images are processed into a single perceptual image by determining the difference (decomposing the difference by determining the box-like depth). merge into.

Three-dimensional reproduction devices are a contradiction in terms, such as movies, television, or any device that attempts to reproduce dimensional images on a flat screen.

Accurate reproduction with the same properties is that the convergence of the stereo camera is such that when the observer shifts his line of sight from the foreground to the background,
It is interesting to note that it is a priori synchronized with the convergence of the observer's eyes. Actually, in certain complicated scenes in existing Ujigen movies,
There is a double image problem.

This order of magnitude is not as obvious as using special glasses to separate the stereo images in a dimensional movie, but it is still present.

In the display or projection of moving video scenes, such as in movies and television, the video update rate (60 ft/l/t for television, 〇l gnome)
It is well known that this is sufficiently larger than the critical flicker frequency to create the illusion of a continuously moving image. However, the stereo partner according to the present invention (one of the pair)
The alternating speed of ] works well at low alternating speeds near or below the critical flicker frequency. This may be because the lower speed can cause binocular competition of neural cells in the visual cortex, resulting in depth perception. If the stereo object images that are alternately displayed at low speeds are made to overlap appropriately in this way, no alternating video speed will be recognized and no flicker will occur. However, when the partner object image exceeds the limit of binocular fusion, the partner image flies or flickers at the partner alternating speed. Increasing the partner alternating speed reduces flicker, but a continuous double image occurs for objects outside the limits of binocular fusion, and a double image occurs for objects within the limits of binocular fusion. The dimensional illusion becomes smaller, resulting in a continuous, blurred object.

In the case of dimensional reconstruction devices, it seems impossible to keep all corresponding elements of a complex scene within the limits of binocular fusion for all observers. When a person shifts their gaze from a foreground object to a background object in a real dimensional scene, the angle of convergence of the eye changes, so that, as shown in Figures 1B and 1C,
Change the relative positions of the foreground and background object images. However, 5
In a dimensional reproduction device, the observer's eyes are focused on the screen. The convergence angle of a viewer's eyes can change when the viewer's line of sight changes from the foreground to the background in a stereo playback scene. However, by careful geometric projection analysis,
Since the camera is primarily focused only on the object of interest (e.g. the foreground), when the observer's gaze shifts to the background object, the correct retinal superposition of the background object image creates an unnatural convergence angle for both eyes. It became clear that the request was made.

[Means for solving problems]

The present invention is designed to ensure that all necessary objects in the stereoscopic reconstruction are within the binocular fusion limits of the observer, allowing binocular fusion to occur without foreground blur or splash effects in the background. The present invention provides a video processing device that processes stereo video signals. This can be accomplished in one of two general ways. One method is to alternately provide stereo video only from objects that are within the limits of binocular fusion. For objects that are outside the binocular fusion area, only the video from one of the two cameras is passed through, but that video is always passed through. The second method is to create a stereoscopic image of each object in the scene from all necessary objects such that both background and foreground are arranged so that binocular alignment occurs (as illustrated in Figure 1d). Reorder the video.

According to the invention, a stereoscopic lantern pair of images is presented, one partner at a time, but in such a way that the two images are presented alternately so that each partner is exposed to both eyes simultaneously. . The alternating speed is on the order of 3 to 25 per second for each video, but may be varied for optimal effect. The relative exposure times for each partner may be equal or
It may be changed to obtain the optimum effect.

The "single layered video technique" produces vergence binocular video combined with monocular video to eliminate the limitations of the binocular fusion area.

For detecting corresponding elements in two electronic video signals from dual video cameras and when corresponding elements of the two electronic video signals cause the reproduced stereoscopic elements to exceed the limits of binocular fusion of the observer. Additional techniques are provided to detect failures.

In these techniques, corresponding stereoscopic elements that are within the binocular fusion limits of the observer are alternately presented, but when the corresponding elements are outside the binocular fusion limits, only one side of the stereo video signal for that element is emitted. are coupled to a method and apparatus for combining parts of each separate stereoscopic image by passing the stereoscopic images continuously and without alternation. Three-dimensional elements that are shown alternately appear to have depth, while elements that are not evenly spaced alternately appear to have depth (-1
DA (two) can be seen as foreground) elements.

Single binocular fusion with a three-dimensional illusion angle when all the necessary video information from a scene (foreground and background) is in the correct binocular relationship and the two stereo partners are swapped at the appropriate partner alternating speed. Additional techniques are provided for rearranging the video information in either one or both stereo electronic video signals such that a stereo electronic video signal is perceived. Regarding these techniques,
Corresponding image elements that would otherwise be outside the binocular fusion limits are placed so that they are within the binocular fusion limits.

The techniques described above may use one conventional horizontally scanned television camera and a horizontally spaced stereo camera. Additional techniques are provided including vertically separated stereo cameras and vertical scan television cameras. These techniques can significantly reduce the video signal processing required to achieve globally focused stereo video. (For purposes of this discussion, the term "globally focused" stereo video means stereo video in which all corresponding elements are within the limits of binocular fusion throughout the reconstructed image. Another object of the present invention is to create dimensional images that can be adapted for artificial image creation for use in television broadcasts (standard and non-standard), closed circuit television, and video games, animated cartoons, television cartoons, and similar applications. This is the creation of a method and device for nodame.

〔Example〕

FIG. 1A is a plan view showing the line of sight from the observer and each eye to a foreground object F and a background object B to help explain the invention.

FIG. 1B is a schematic representation of the superimposed retinal image for the viewer of FIG. 1A when the viewer's eyes are focused on a foreground object F. FIG.

FIG. 1C is a schematic representation of the superimposed retinal image for the viewer of FIG. 1A when the viewer's eyes are focused on background object B. FIG.

Figure 1D simulates a video processed according to the present invention so that the observer's eyes can focus on both object F and object B simultaneously, creating a binocular fusion of both objects. FIG. 2 is a schematic diagram of a superimposed retinal image when

Referring to FIG. 2, two television cameras 21o and 212 with identical opto-mechanical components are properly spaced and aimed at a target to capture video in stereo relationship to each other. Two cameras are shown focused on object F. Video switch 211
+i, In response to a command from the control device 216 that controls the switch 214, the video can be passed from one camera and then from the other camera, and the video synchronization generator 21g is connected to the control device 216. Sync your camera's video. Transmission medium 220 is switch 2111
The selected video is sent to the monitor 222.

Although the embodiment of FIG. 2 is not specifically designed to reduce or eliminate the limitations of the binocular fusion problem, it may be used where all video elements are at relatively equal distances from the two cameras. I can do it.

The two cameras 210 and 212 are focused on the desired object, and the video signals from the two cameras are synchronized by a synchronization generator 2111+. 2
Video from the two cameras is applied to video switch 21+1. The control device 216 receives a synchronization signal from the synchronization generator 21g. Controller 216 controls switch 2111'i so that video can be passed first from one camera and then from the other camera. Video from two cameras is approximately 3 to 2 seconds per second for each video.
5 speeds may be applied alternately and adjusted for optimal effect. Each channel may be exposed for the same amount of time.

However, the relative exposure times may be varied to obtain optimal effects, and the technique of US Pat. No. 11,1006,291 may be applied to the present invention.

Experiments have shown that the optimal alternating speed may be a function of the light intensity of the video scene. This result was published by Robert Efron in Pritish Jernel op Ophthalmology, published December 195.
This is supported by the paper ``Stereoscopic vision'' by Fron (3). Therefore, by using a photometer to measure the total light intensity of the video scene and operating it by connecting the output of the photometer to a circuit that controls the alternating speed of the stereo video signal, The alternating speed can be slightly optimized. An alternative method is to electronically measure the amplitude of the video signal, which amplitude is a function of the light intensity.

Layered Video Techniques Layered video techniques are generally applicable when each part of a video scene can be photographed separately and combined by the 1% special effect technique commonly used in the video industry. For example, a singer can be photographed against a uniformly colored background to create a focused binocular video of the singer. You can take photos of the orchestra separately with a monocular video. Existing special effects circuitry can then be used to detect and replace the uniformly colored background with overlapping video. This method focuses the image of the singer on the background as a monochrome video with the depth effect of a binocular video.
The overall result is that there is no flicker or jumping in any part of the video.

Many variations of this technique are possible. For example, in the previous example, the orchestra could be photographed with a stereo camera (but a different stereo set than the camera for the singers) and combined with a stereo image of the singers. Additional layers of video can also be combined. Several binocular videos can be combined in real time, or videotape recording can be used.

However, everything must be properly synchronized and focused.

Correspondence Detection Technique As the television camera scans the scene horizontally as a function of time, the relative time of occurrence of the corresponding stereo video elements within the horizontal scan is used to ensure that each pair of corresponding stereo elements is within the limits of binocular fusion. You can decide whether This correspondence detection can be used to perform alternation of focused stereo video elements while maintaining continuous video from one camera for unfocused video. (For discussion here.

A "video element" may be loosely defined as a video signal from the smallest discernible portion of a single object within a video scene. Corresponding video elements are video elements in two stereo video signals that originate from the same small portion of the same object in the video scene. ) The first step in determining whether the corresponding video elements are within the limits of binocular fusion is to identify the corresponding elements by the correlation one-leg technique. Correlation measurements can be achieved in various ways. Correlation, amplitude comparison between two video signals (Figure 3), slope comparison between two video signals (Figure 4), correspondence measurement between two signals (Figure 5)! , or other time domain techniques. Frequency domain techniques can also be used. That is, by detecting corresponding frequencies in two signals. It is often necessary to combine the techniques described above. Correlation measurements can be accomplished with monochromatic signals, color signal mixing, or a combination thereof. Correlation determination can be accomplished by analog techniques, or the video signal can be digitized and digital techniques used. A powerful combination technique for color signal correlation measurements is the combination of light intensity signals and color signals. Correlation measurements can be accomplished in the vertical direction, in the horizontal direction, or in the vertical and horizontal stitching directions. The two directional techniques can be easily developed using digital techniques and algorithms for image processing using digital computers.

A logic circuit designed to indicate that each element corresponds is used to detect when two video elements in two stereo video signals are sufficiently correlated.

FIG. 3 is a schematic diagram of an amplitude correlation measuring device for determining the amplitude difference between two stereo video signals, Video 1 and Video 2. Video differential amplifier 10 amplifies the difference between the two video signals. Threshold (gain) adjustment device 314
(When video 1 is a suitable value higher than video 2, the output of amplifier 310 is a logical sum) 3111! is used to adjust the gain of amplifier 310 to force the output of 1 to a "1" state indicating that the signals are not properly correlated. Amplifier 312 is the same as amplifier 310 but connected to the two video signals with opposite polarity.

Therefore, when the amplitude of video 1 is a threshold above or below that of video 2, the OR gate will produce a ``1'' indicating uncorrelation.
Output. If the two video signals are approximately equal, the OR gate outputs a "0" indicating correlation.

FIG. 4 is a schematic diagram of a slope correlation measuring device for measuring the total difference in slopes of two video signals. Reference numbers 1110 and l
112 is a video amplifier connected as a differentiator whose output is proportional to the slope of the video signal input. Differential amplifiers l116 and 111g, threshold adjusters 1L20 and 1122 and OR gate lI2! works in the same way as in Fig. Therefore, when the slopes of the two signals are approximately equal, the OR gate +1211 outputs a ``0'' indicating that the slopes are properly correlated. If not, it indicates that the signals are not properly correlated.
1" is output.

(U 8) FIG. 5 is a schematic diagram of a video edge detector.

A video signal is connected directly to one terminal of a differential video amplifier 512 and through a heavy-duty delay line device 510 (eg, a video sampler/hold circuit) to the other terminal. The amount of delay is shown as 200 nanoseconds, but may be changed for optimal results. video signal is 2
If the threshold changes by some threshold within 0.00 nanoseconds, the output of amplifier 512 causes digital counter or flip-flop 516 to change state indicating that a video edge has occurred. A threshold adjuster 51+1 is provided to adjust the amount of change in the video signal necessary to represent an edge as required by the application.

Next, an example of the corresponding detector will be explained for a simple video scene. With full reference to FIG. 6, correctly synchronized left and right stereo video inputs are simultaneously provided to edge detectors 606 and 608 and sample/hold circuits 610 and 612. Each sample/hold circuit N6to and 612 samples the signal immediately after an edge (eg, λ - 200 nanosecond sample window) and holds the respective signal for input to amplitude correlation detector 6111. Amplitude correlation detector 611
1 compares the amplitudes of two sampled signals and outputs "1" when the signals are properly correlated, and outputs rOJ when the signals are not properly correlated. Properly correlated signals can be said to be "corresponding."

It is clear that an important part of correspondence identification and detection is the sensing or detection of the edges of various objects in the video scene. Special techniques can be used to fully identify the edges of a video object to separate various foreground, midground, and background elements. for example.

Foreground objects can be illuminated from behind with ultraviolet or infrared light. The signals from the 5 special video sensors can be scanned in a step-by-step manner with all regular video sensing elements and a special "videocon" sensitive to backlight. Helps identify edges of video objects.

"Distance gating" can also be used to identify edges of video elements as a function of distance from the camera. The amount of video excursion required to enable binocular fusion is a function of the distance of the video object from the camera; however, the distance is not typically measured by the camera. Radar devices can measure distance by transmitting short pulses of energy. Since the pulse travels at a characteristic velocity, the distance to an object can be calculated from the time it takes for the pulse echo to return from the object. Radar range at an object is measured by taking short successive time samples (range gates) of the reflected energy to determine the time of arrival of the reflected energy from the object. This principle can also be used to implement edge detection for various objects as a function of distance to the object, such that the video deflection required for binocular fusion is a function of distance at video object 1.

and 'n can be realized in video equipment by transmitting pulses of light and by recording the scene for short successive periods of time. Each time sequential "Note"
Since it contains videos from objects at successively increasing distances, it can be used to implement distance-dependent video excursions.

It can be realized as a single laser that converts all optical pulses.

Alternatively, it may be an ordinary optical strobe, or it may be embodied in a mechanical device such as a rotating mirror used in very high speed cameras. Light is 300m/μse
Since the pulses must be short in duration, on the order of less than 1 microsecond, the exact time of transmission must be well controlled. The sequential sample time must also be well controlled, on the order of 20 nanoseconds. The light pulse may be in the invisible part of the spectrum (ultraviolet or infrared) and may only be visible by the sensor that uses it. One embodiment of such a device is shown in FIG. As before, two optically related stereoscopic cameras 710 and 712 are used. But 5
The camera 712 has a prismatic mirror 716 that separates incident light between a normal camera sensor 7111 and a sensor 71g that is sensitive to light from a flash light source 726 (+12). Flash light source 726
is strobed several times per second, emitting short bursts of light on the order of one microsecond duration. The emitted light is reflected from the video scene and focused by camera optics and prism 716 onto light sensing element 718 .

Sensor 71F3 may be several elements or may be a single element that sequentially samples the emitted light at successive time intervals. During the first time interval, from 10m to 10m
Identify video elements at a distance of .

then also identify video elements at a distance of 5 to 20 meters;
...You may close it. Controller 720 controls video processing unit 722 to identify the edges of the elements and temporally shift the appropriate video elements in the regular video from camera 712 to cut the two videos as described above. The information from sensor 71g is used to ensure that the depth illusion occurs without the limitations of binocular fusion problems when combined with the video from camera 710 by changing υ.

The distance to a certain object can also be determined by the Sankaku-Oki 1 quantity. This can be achieved by providing a second camera on one layer of the stereo camera. If this camera, along with the camera below it, is focused on an object, the corresponding video elements from that object will correspond vertically to exactly the number of horizontal scan lines within the horizontal scan range. Become. Corresponding video elements from other objects at different distances are at the same position in the horizontal scan, but those for the upper camera occur on different horizontal scan lines than for the lower camera. Video from one of the two cameras may be passed through a delay line (eg, Fairchild Semiconductor Part No. CCD321A1 Broadcast Quality Video Delay a) equal to an integer multiple of the horizontal scan time. The signals from the two cameras are then processed by a correspondence detection circuit to determine in which horizon the corresponding elements occur, allowing the circuit to calculate the distance of each element from the camera. This information may be used to implement extensive video selection techniques, and may also be used in comprehensive video processing techniques.

The relative times of occurrence (within one horizontal scan) of vergence detector-enabled video elements can be used to determine whether those elements are within the limits of binocular fusion. The correspondence detector discussed in the previous section may be used to detect corresponding video elements in two stereo video signals. A device designed to detect when the corresponding video element is within the limits of binocular fusion is a radial isthmus detector. The radiation detector is designed to receive the signal from the edge detector and the signal from the corresponding detector. Referring to FIG. 1C, when an object is closer to the camera than the distance from the convergence point to the camera, the right camera video from that object is further ahead than the corresponding left camera video. Similarly, when the object is after the convergence point (1B
The left camera's video from the object is ahead of the right camera's video when the image is scanned from left to right.
15') Measure the time leading or lagging behind the corresponding video element from the camera. If the lead or lag time is p less than some (experimentally determined) threshold, then those video elements are y.

They are shown to be (as appropriate) converging so as to obtain ocular fusion; otherwise they are defined as being outside the limits of binocular fusion.

The vergence video selection technique can also be used to overcome the limitations of the binocular fusion problem from five-dimensional video. Such a device is shown in FIG. 8, which shows switch F! ill in a different way, i.e. the light detector 83
Follows the basic technology except that it is operated in conjunction with 2
810 and 812 are shown.

Representative signals from the two cameras 810 and 812 of FIG. 8, where the cameras are focused on object F, are shown in FIG. Correlation, correspondence and timing circuitry in vergence detector 830 identifies corresponding elements in the two video signals and determines the relative occurrence of corresponding image elements, shown as times T1 and T2 in FIG. Set the time. In this case, time T1 is within the time threshold for binocular fusion, and time T2 is greater than that threshold. Therefore, this circuit does not switch and the switch output in FIG. After a short period of time, the circuit passes the video from the left camera, denoted as is outside the limits of binocular fusion), pass the video denoted as switch output ≠ 2.In this way, the video of object F, which is within the limits of binocular fusion, is
The threads are alternately threaded at a speed of 25 to 25 turns per second, as previously set. However, videos from unobtrusive objects, such as object B in the examples of FIGS. 8 and 9, which may cause binocular fusion problems, cannot be passed alternately. In summary, only videos that are within the limits of binocular fusion are presented at the appropriate speed and for the appropriate portion of the horizontal scan, alternating for all scans within each frame. The other video is fed continuously from one of the two cameras.

Switching circuits for combining two video signals are already available in various devices in the television industry, such as "special effects" products. A gated video amplifier such as Motorola part number MC111115 may be used.

One variation of the light chain video selection technique is one master.”
The solution is to use a camera and several partner stereo cameras. For example, a first of one partner stereo camera may be focused on the foreground object together with the master camera, a second camera may be focused on the master camera and the midground object, and a fifth camera may be focused on the master camera and the midground object. Concentrate on the master camera and background objects. Multiple sets of vergence detectors and switches would be required. The first set is connected to the master camera video and the first partner stereo camera to select and alternately send foreground object video with the master camera video. The second set is connected to the master camera video and the second partner stereo camera to select and alternately send the mid-ground object video.

The fifth glue works similarly (the order in which the videos are mixed or switched is designed such that the foreground object video replaces the middle ground video being superimposed (e.g. the background video is mixed or alternated first, then the middle ground video is superimposed). (middleground, then foreground, and so on). In this way, several "layers" of video can be processed at once to obtain binocular fusion. Another variation that can be achieved with just one is the following scenario. Two cameras are aligned to the foreground object. The right camera is used as the master camera, and the foreground object video from the left camera's video is The video from the left camera is delayed for a short period of time (e.g., 300 nanoseconds). A second set of light chain detection and switching circuits brings the left camera video from the middle ground, i.e. the next silk object, into the correct time and position relationship (relative to the right camera video) required for binocular fusion. is used to switch the left camera's middle object video with the right camera's video. Delay, (9) Apply this sequence of light chain detection and camera video alternation to the middle object video at mid distance depending on the light needs. It can be repeated several times on the object.

Light Chain Processing Techniques The previous section describes how to select and alternately send corresponding video elements from objects that are within the limits of binocular fusion. - Describe video processing techniques in which video elements are temporally shifted to be within the limits of binocular fusion.

Referring to FIG. 10, left video signal 110 and right video signal 112 are produced according to the basic technique described above.

Corresponding detector 10111 operates in the same manner as described above in connection with FIG. 6, except that it generates control signals that control left and right video processing procedures 1016 and 1018. Video processing units 1016 and 101g process video elements of one or both of the video signals in stereo relationship such that the various corresponding video elements are in convergent binocular relationship to each other necessary for binocular fusion. Shift to approx. The two video signals can then be sent alternately in the manner described above to create the illusion of full depth. Time shifting can be accomplished using the analog shift register capabilities of an analog device (e.g. Eva Fairchild Semiconductor part number CCD32181 or similar device), or the video signal can be digitized to create a time shift or delay. This can be achieved using a digital storage device as an intermediate storage device.

The analog version of the video processing device can be implemented as follows. The two stereo cameras focus on the background such that all corresponding background video elements from the two cameras are aligned in time. Video elements in the right camera's video from medium volume and foreground objects lead the corresponding elements in the left camera's video. At the beginning of each horizontal scan, video from both cameras is passed through without delay. In this implementation, the left camera video is always passed through without delay. The right camera video is also passed through at a point where the corresponding detector detects a non-correspondence between the two video signals, and then the right camera video is slowed down at the point where its corresponding video element appears in the left camera video. I was forced to do it.

Both signals are then passed. When the right camera video is delayed, control signal 1022 controls switch 1020 to overlay the left camera video with the right camera video. The timing signal controls the switch to capture the left camera video and the right camera video processed in the manner described above at 3?
The number of alternations is l, -=L25.

A five-dimensional television device can also be realized using a digital computer. If you look at Figure 11.

Video L and video R are generated from two cameras in a stereo relationship. The video may be monochrome or a combination of monochrome and color signals depending on the requirements of the application. Timing device 1116 = AD converters 1110 and 111
2 to digitize the video at a rate suitable for proper signal reproduction (on the order of a few megahertz). The digitized signals from AD converters 1310 and 1112 are
The data is transferred to an input digital storage device 111 which is under the control of a digital timing device 1116. The sample resolution (number of focus points) of the sample processing must be adequate to create the desired effect. The computer's software is designed to control the computer 111 B''fx to examine each digitized video element individually and sequentially as necessary to identify each object element within the video scene. For example.

A plain object has equal video words for each adjacent video sample of the object. Equivalent words can be identified by the computer and the size of the object can be determined in horizontal and vertical position relative to both the left and right video cameras. When the corresponding videos from the two cameras are not in a straight video relationship, the computer software is designed to horizontally shift one or both of the digital representations of the object so that the video object is visible. . The background video from the other stereo image can be used to "fit" objects to smooth edges.

Computer 111g can then output the samples to output digital storage 1120 in the appropriate order to produce two child stereo images at the alternating speeds described above in the basic technique. Digital bitio is DA converter 1122
is converted back to analog form by The DA output can be filtered to reconstruct the synchronization signal and other signal processing techniques can be applied as appropriate.

The result is a stereo video image in which all objects in the image have appropriate alternating binocular disparity, but each corresponding binocular object is appropriately light chained to allow binocular fusion to occur. The distortion reproduces an image with the illusion of depth but without the flicker and jumps caused by the limitations of the binocular fusion problem.

The methods described above may require that stereo video be recorded so that it can be processed at a slower rate than its actual temporal rate. Special purpose circuits with "1-dwired" algorithms can be used for real-time processing in certain applications.

In vergence processing techniques, when shifting the position of a video object, it is necessary to preserve the object's size, shape, color, and texture (51+). When the edges of hidden midground objects on one of the two stereo screens are evened out by shifting the foreground objects, that screen is filled with the stereo elements of its partner. It is not essential to alternate the background elements since the difference is not normally perceived as background.

Although the embodiment described herein is for an alternating stereo video system, other systems, such as using pentachrome or polarized glasses, may process the video as described herein to obtain the overall stereo image. There will be a marked improvement when providing light chains.

The visual perceptual response to images produced by horizontally spaced stereo cameras is well known.

The illusion of depth also occurs when the camera is moved vertically away. A three-dimensional television apparatus according to the invention using vertically spaced (horizontal scan) cameras can be constructed using the apparatus of FIG. 2, except that one camera is placed on top of the other. It can be realized in the same way as the device described above.

The camera separation should be approximately the same as before for horizontally spaced cameras. This visual perceptual device may have a vertical disparity fusion capability that is less than the horizontal disparity fusion capability, so it may be necessary to reduce the separation distance slightly. Switching of the two video signals is accomplished as before at a rate of 3 to 25 seconds per partner for each partner. The camera may be focused in the same way as the bin) and the relative exposure times may be varied to obtain the optimum effect, as described in US Pat. Apply the technology of oo62 item i. The vertical separation of the camera is described in U.S. Patent No. 3. It is also applicable to the alternating interlaced field technique described in 'l 57.3611.

When the camera is moved vertically away, corresponding video elements in the two stereo signals may occur at the same point in the horizontal scan line of the screen, but on different scan lines.

A block diagram of a video light chain processor designed for vertically spaced cameras is shown in Figure 12A.

For a camera focused on the background, the video elements from the lower cameras are evenly spaced to the same scan line or partial scan lines as the scan lines of the corresponding video elements of the upper camera. 12th A
The illustrated design assumes that the geometry of the video scene is such that each video element is within five lines of its corresponding stereo video element, but the design can be expanded as the application requires. Figure 12B shows the same block 1216.1218.1220.1222.12214 in Figure 12A.
and 1226 additional details.

The purpose of the apparatus of FIG. 12A is to replace each video element of upper camera signal 1210 (which element was not replaced by an element of signal 1212) with a corresponding video element from lower camera signal 1212, and then Alternating at the aforementioned speed of 3 to 25 impacts per second. A fully convergent binocular video results.

To accomplish this, the video signal 1210 from the top camera and the video signal 1212 from the bottom camera are input to a correlation detection and switching block 1216, the details of which are shown in FIG. 12B. There is. Logic "0"
” is added to both inputs of the OR game) 1272 (only for blocks 1216 and 121g). Timing signal 1261+ transfers these signals to sample and hold circuits 1266 and 1268 and flip-flop 1.
27+1 as a clock signal. If the outputs from sample and hold circuits 1266 and 1268 are properly correlated and the signal from flip-flop 12711 is "0", signal 1252 from correlation detector 1276 is sent to switch 1270 to sample and hold circuit 126g. Pass through the lower camera's video i6. Otherwise, the upper camera video is not passed.

Next block 1220.1222.12211 and 1
Two inputs 122 to OR gate 1272 at 226
The purpose of g and 1230 is (1) to avoid replacing again the video element of the upper camera signal that has already been replaced, and (2) to replace the video element of the lower camera signal with the video element of one of the upper camera signals. Elements should not be replaced.

The signal 1232 generates a control signal 14 to achieve the 5-th round in the i next stage. Correlation detector 1276 connects switch 12
Passing the lower camera video element to 70 also emits logic rlJ=e on signal 12. The output from block 121g takes two paths to block 121g.
20 and shift register 12+12. The path to block 1220 is function (1) above 5.
, and the path to shift register 1242 is function (2
Achieve 1. Shift register 1212 provides an appropriate delay to keep the control focus in synchronization with its associated video element when the control focus is delayed by delay element 121411.

CD/SW block 1216 replaces all background video elements. Block 121g replaces the video elements from the background, 7) the immediately preceding object; Block 1218 is connected to replace the upper camera video with the lower camera video from the adjacent scan line of the next interlaced field so that vertical delay block 123+1 is approximately 1
It is necessary to delay the lower camera video by 60 seconds.

Block 1220 replaces the video element from the next closest object occurring in the next horizontal scan line. Horizontal delay element 12110
delays the bottom camera video for one horizontal scan interval to make the bottom camera video coincide in time with the two scan lines of the top camera video below it.

This dissolution is repeated with the top camera video being processed with the same horizontal scan line and five horizontal scan lines above that of the bottom camera video.

Delay element 1254 delays the unaltered upper camera video by an appropriate amount to allow the two signals to enter switch 1256 simultaneously. Synchronizer 1262 and controller 1260 provide signals to switch 1256 for alternating the two videos to each extensor muscle from 3 to 25 seconds to obtain a fully saturated alternating binocular video.

The apparatus of FIG. 12A may be implemented in analog form, or the unselected video signal may be digitized, or the above-described apparatus may be implemented with digital circuitry. In the case of analog methods,
The delay element was replaced with the aforementioned Fairchild part number CCD32.
It can be realized with 1A1.

Digital methods use digital storage as a delay element.

Vertical scanning horizontally spaced stereo camera The above-mentioned vertically spaced camera system uses light-chained stereo cameras as a whole.
It is a particularly reliable and easily implemented method of making video, and can be implemented with a standard horizontal scan camera. but,
The 5-dimensional effect from vertical disparity may be worse than horizontal disparity. A similar method but using horizontal disparity can be used with a vertical scanning camera. The apparatus of FIG. 12A operates correctly as long as the camera separation is perpendicular to the camera scanning direction.

A five-dimensional television device according to the invention and using a vertical scan television camera may be implemented as shown below. This apparatus should correspond to that described above for the apparatus of FIG. 2, except that a vertical scan television camera is used. For cameras scanning down from the ground and interlaced scan lines going from left to right, the left camera video should be connected to video input 1210 in the twelfth diagram. Right camera video is video input 121
Should be connected to 2. The apparatus of FIG. 12A operates as described above.

The overall light chain binocular video thus produced is mounted on standard broadcast equipment and standard television receivers at 90° from perpendicular.
°Create a defeated image. However, it can be converted to a video signal that is compatible with a scan converter consisting of a series of video storage elements, such as the Fairchild part number CCD321A1. Each scan line of video is buffered within the storage element until a complete frame is stored. The video is scanned through with one video element from each scan line going from left to right. That is, one horizontal line of the video is repeated five times, first the top line, then the second line, then the fifth line, and so on until the entire video is output. . Alternate video lines are delayed for about 1/60 of a second to provide a standard interlaced field. It may be necessary to have two storage arrays so that the video can be stored in one array and previous frames processed from the other array. A vertical scan camera is a high resolution camera that can ensure that the video it produces is of broadcast quality.

FIG. 13 shows an implementation of a layered video device designed to produce totally convergent binocular video.

Such a device includes one foreground camera] 12 and 1
Background 131 with a uniform color and approximately equidistant from 3111
A separate stereo video image of only the foreground object 1310 in front of 6 is created. The middle ground object 131g is the second pair of cameras 1
Plain background 13 approximately equidistant from 320 and 122
It's in front of 211.

A separate monocular video camera 126 produces a monocular video image of only the background object 132g. Background camera 132g.

Mid-range cameras 1318 and 1520 and foreground camera 131
Video images coming from cameras 131 and 1312 are combined under the control of switches 1330 and 1332 and synchronizers 13 to 4 as described above in connection with FIG.
The solid background of the mid-ground video from the background camera 1328 is replaced by the corresponding portion of the monocular background object video from the background camera 1328, and the solid background of the foreground video from the foreground cameras 1312 and 1 is replaced with the mid-ground. The background is to be replaced by the corresponding part of the combined video.

[Brief explanation of drawings]

1A to 1D are diagrams illustrating the geometry of binocular viewing, and FIG. 2 is a block diagram of a basic embodiment of a five-dimensional television system according to the present invention. FIG. 5 is a logic circuit diagram illustrating amplitude correlation measurement. FIG. 4 is a logic circuit diagram illustrating slope correlation measurement. FIG. 5 is a logic circuit diagram illustrating a video edge detection device. FIG. 6 is a block diagram of an apparatus for detecting when video elements in two stereo video signals correspond; FIG. 7 is a block diagram of a pulsed optical distance measurement and convergence apparatus; FIG. 8 is a binocular FIG. 9 is a block diagram of an apparatus for producing monocular video for objects in extension, alternately transmitting stereo video elements only for objects that are within the limits of fusion; FIG. 9 shows a signal illustrating the alternation of focused video elements; FIG. 10 is a block diagram of an apparatus for processing stereo video to bring all corresponding video object elements within the limits of binocular fusion. FIG. 11 is a block diagram of a stereo video processing apparatus that produces alternating stereo video such that all corresponding video object elements are within the limits of binocular fusion. 12A and 12B are block diagrams of a stereo video processing apparatus for vertically spaced stereo television cameras, and are also applicable to horizontally spaced vertical scan television cameras; FIG. 13 is a block diagram of a stereo video processing apparatus for a vertically spaced stereo television camera; FIG. 1 is a block diagram of a device according to the invention that uses layered video techniques to create an alternating signal; FIG.

Claims (1)

Claims: 1. A method for displaying a stereo pair of images so as to present a single three-dimensional, sharply focused, flicker-free image to human visual perception, comprising: (a) a first pair of images; (b) creating separate stereo video images of only foreground objects placed approximately equidistant from the camera against a plain background; and (b) a second pair of cameras positioned approximately equidistant from the camera against a plain background. (c) separately creating a monocular video of only background objects; (d) converting the plain background of the midground video into a monocular background; combining the foreground, midground and background video images such that a corresponding portion of the object video is replaced and a solid background of the foreground video is replaced by a corresponding portion of the combined midground and background video; ) displaying a first combined video signal of the signal from the background video and the signal from one camera of each pair of foreground and midground camera pairs; and superimposing a second combined video signal of the signals from the other camera of the mid-view camera pair with the display of the first signal and displaying the first signal in binocular relation for approximately the same amount of time as the first signal; , (g) switching between the display of the first and second signals at a rate of between 3 and 25 per second for each image. 2. A pair of video cameras viewing a scene in a manner that displays stereo pairs of images to present a single three-dimensional, sharply focused, flicker-free image to human visual perception. B. displaying the video signal from the first camera; C. displaying the video signal from the second camera in stereo relationship with each other; (1) (a) displaying the video signal from the second camera; (b) comparing the characteristics and temporal occurrence of video elements in each of the two video signals; (c) the two video signals that are not within the limits of human binocular fusion; (d) processing the video signal from one camera to time shift the video elements outside the limits of binocular fusion, and the processed video signal is used for the switching described below. bringing such video element within the limits of binocular fusion when displayed together with a video signal from a second camera according to the signal, in binocular relation to the first signal by (ii) displaying and superimposing the signals, (iii) displaying the first signal for approximately the same amount of time as the first signal; A three-dimensional television system that consists of a process of switching at a speed of 3. The three-dimensional telecasting system of claim 2, wherein said comparing step is performed by detecting edges of video objects and comparing characteristics of video elements within such objects. 4. The identification step is carried out by illuminating some objects in the video scene from the background with light invisible to the video camera and detecting elements of the objects so illuminated from behind. The three-dimensional television system according to scope 2. 5. A three-dimensional television system according to claim 2, wherein said step of identifying is performed by distance gating using pulsed electromagnetic radiation. 6. The three-dimensional television system of claim 2, wherein said video signal processing step includes time delay shifting of at least two video elements of one of the video signals by different delay periods. 7. A method of displaying stereo pairs of images so as to present a single three-dimensional, sharply focused, flicker-free image to human visual perception, including (a) a pair of video cameras that capture a scene; (b) locating video elements in the scene that are not within the limits of binocular fusion and video elements that are within the limits of binocular fusion for viewing; (c) from a pair of cameras such that video elements within the limits of binocular fusion are displayed alternately for approximately equal times at a rate of between 3 and 25 per second for each image; (d) continuously displaying video signals from only one of the pair of cameras in place of video elements that are not within the limits of binocular fusion. John method. 8. (a) A pair of video cameras viewing a scene in a manner that displays stereo pairs of images so as to present a single three-dimensional, sharply focused, flicker-free image to human visual perception. (b) displaying a first video signal from one of the first cameras; and (c) displaying a second video signal from a second camera in a stereo relationship with respect to each other. (d) displaying the first video signal in binocular relation to the first video signal for substantially the same amount of time as the first video signal; (e) monitoring the light level of the scene and switching between the first video signal and the second video signal in response to higher light levels; A three-dimensional television system comprising: a step of increasing the switching speed between; 9. A method of displaying stereo pairs of images so as to present a single three-dimensional, sharply focused, flicker-free image to human visual perception, comprising: (a) displaying multiple pairs of images to view a single scene; (b) arranging video cameras such that each pair of cameras is in stereo relationship with each other and each pair of cameras is focused at a different distance within the scene; (b) from one camera in each pair of cameras; (c) displaying the first video signal; (e) combining a first composite video signal depicting the scene from one of the cameras using a congestion detection circuit, a switching circuit and a layered video circuit; (f) between the display of the first and second composite video signals; a three-dimensional television system comprising: switching for each image at a rate between 3 and 25 per second; 10. A method of displaying a stereo pair of images so as to present a single three-dimensional, sharply focused, flicker-free image to human visual perception (a) a pair of videos, one above the other; (b) displaying a first video signal from one of the first cameras; and (c) a second video signal from a second camera; (d) superimposing a display of the second video signal with a display of the first video signal and displaying it in binocular relation to the first video signal for approximately the same amount of time as the first video signal; switching between the signal and the second video signal for each image at a rate between 3 and 25 per second. 11. (a) A pair of vertical scanning video cameras are used to display a single scene in a manner that displays stereo pairs of images so as to present a single three-dimensional, sharply focused, flicker-free image to human visual perception. (b) displaying a first video signal from one of the first cameras; and (c) displaying a second video signal from a second camera. (d) displaying the first video signal in binocular relation to the first video signal for approximately the same amount of time as the first video signal; A three-dimensional television system comprising: switching between video signals for each image at a rate between 3 and 25 per second. 12. (a) A pair of video cameras viewing a scene in a manner that displays stereo pairs of images so as to present a single three-dimensional, sharply focused, flicker-free image to human visual perception. (b) displaying a first video signal from one of the first cameras; and (c) displaying a second video signal from a second camera in a stereo relationship with respect to each other. (d) displaying the first video signal in binocular relation to the first video signal for a significantly shorter period of time than the display of the first video signal; A three-dimensional television system comprising: switching between video signals for each image at a rate between 3 and 25 per second. 13. A television device that displays a three-dimensional, sharply focused, flicker-free image to human visual perception, comprising: (a) a pair of video cameras arranged in stereo relation to each other to view a single scene; and (b) transmitting a video signal from a first camera of said pair of cameras and then transmitting a video signal from the other camera for a substantially equal period of time from the first camera and the second camera. (c) means for receiving the transmitted video signals and displaying the signals in superimposed binocular relationship with each other; and the means for displaying in binocular relation comprises: (i) means for measuring characteristics of each of the two video signals; and (ii) means for measuring characteristics and temporal occurrence of video elements in each of the two video signals. (iii) means for identifying corresponding video elements in each of the two video signals that are not within the limits of human binocular fusion; and (iv) processing the video signal from one camera. to temporally shift video elements that are not within the limits of binocular fusion to bring them within the limits of binocular fusion when the processed video signal is displayed with the video signal from the other camera. A three-dimensional television device characterized by comprising: a means for providing a three-dimensional television; 14, in a way that creates a pair of stereo images whose corresponding objects are within the limits of binocular fusion of the observer: (a) the video at all distances when the two images are superimposed; (b) repositioning corresponding objects in the two images to corresponding positions so that the objects coincide; (b) filling the voids created in each image with the objects using corresponding image elements from the other image; a method comprising: filling by such rearrangement of;