JPS58145291A - Constituting method of stereoscopic picture - Google Patents

Abstract

PURPOSE:To form a stereoscopic image easily, by discriminating a long distance back and forth from small projection and recession in synthesizing a reproducing image and forming a stereoscopic image with both the left and right systems of substantial average original picture such as animation graphic display. CONSTITUTION:In synthesizing the stereoscopic picture by means of animation, two kinds of right and left original pictures 21, 22 corresponding to right and left images 16, 17, are prepared and original pictures 21, 22 are displayed on a monitor TV15 via right and left cameras 11 and 12. The reproducing picture is formed by shifting the right image 16 to the left side of the left image 17 in forming the stereoscopic pictur in front of the screen and by shifting the right image to the right side of the left image 17 in forming the image at the back of the screen. In taking the height of the screen of a receiver for the image forming space as H, the space is taken in a range surrounded with straight lines including three surfaces of a long surface of about 0.39HX0.52H at about 2.7H in front of the screen along the line linking the screen and the observer and a long surface of about 2.8HX3.7H at about 13H at the back of the screen, and the screen itself, allowing to form the stereoscopic picture effectively.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention reproduces on the same screen the visual damage seen from two left and right positions, and the viewer's right eye sees the image from the right position.
The present invention relates to a method of synthesizing a reproduced image in a three-dimensional television apparatus in which the image viewed from the left position is input to the left eye.

It is well known that we are able to see things in three dimensions thanks to our left and right eyes. There are various means for observing stereoscopic images that apply this principle. Recently, 3D televisions for viewing 3D images have been developed by TV M and others. - As an example, we used stereoscopic glasses that were designed so that images taken by two left and right cameras were played back alternately field by field on a receiver's cathode ray tube, and the odd and even fields were input to separate eyes. There is something to use. Stereoscopic glasses use electro-optical shutters such as transparent ceramic batteries.

Recently, advances in computer graphics display technology have made it possible to display moving 3D graphics. Graphic display by computer is designed.

It is used in fields such as design and gaming machines.

On the other hand, in the field of TV comics, animation is gaining popularity, and the emergence of 3D animation is now being considered.

By the way, when you watch these traditional 3D televisions, you can see a lot of hatred jumping out at the viewer from the screen of the machine.
There is no spread to the center of the screen. In other words, after reproduction, the image is normally reproduced only in the space between the screen and the viewer. Moreover, the object that is supposed to be the farthest from the viewer can be observed in the clearest and most in-focus state, which is unnatural.

It is natural for objects that are at an infinite distance to appear blurred and receding into the background of the screen, and objects that are too close to the viewer to appear blurred and moved one step closer to the viewer. It is only natural that only the object that the camera was supposed to capture is clearly reproduced in one screen.

The purpose of the present invention is to overcome the drawbacks of the prior art, and to expand the playback surface of a 3D television to the front side and both sides of the receiver's screen at the center, as well as to the front and back of the screen. Our goal is to provide a technique for reproducing the image in which the image becomes blurred and stutters as you move away from the image, and only the image that appears exactly at the position on the screen shines clearly.

When the angle between the two cameras on a 3D TV is equal to the angle between our eyes and the object, the reproduced image will look natural. In other words, the following relationship is required to obtain a reproduced image as if a real object were located at the position on the screen. First, the distance t between the two cameras and the distance 1lIl between the two cameras and the object! d, from, t.

The angle θ formed by the two cameras is shown by d.

Ru. Assuming that there is a person at the camera position, both eyes are approximately 701 points apart (if the distance between the eyes is t, then the average distance is 1.
It is 0m. ) and the angle θ between the object.

Medium-2, = 70. Here, d is both eyes and d,
d.

It is the distance to the object. The angle θ made by the TV camera is θ
When , the image of me appears as seen with our eyes.

However, since television cameras are generally large, t is 70
■ It is difficult to reduce the size, and the larger the size becomes. In this case, the condition for obtaining natural preparation is d = ratio, and the photograph is taken from a position far away from the object. At this time, the object moves away and becomes smaller, so all you have to do is zoom up and enlarge it by that amount.

In this way, let's consider using a receiver to play back the image of me that we have photographed as if we were looking at an object from the position of the camera. When the distance between the observer and the receiver is D, and the height of the screen of the receiver is H, the clear viewing distance D0 is Do=7H.

In the case of a 20-inch TV, the normal viewing distance is 2,100 cm. Now, the left and right sides of a 3D TV are generally shifted left and right, and the reproduced image appears to protrude from the screen by an amount proportional to this shift. In other words, there is no deviation, the image is reproduced at the position of the matching l1i (#H receiving screen,
A monk who deviates is recognized as a nine-borrow who protrudes in front of the screen by an amount proportional to the deviation. In other words, the larger the deviation, the more the image appears to protrude to the front of the extra plane.

If the amount of this shift becomes too large, it becomes impossible for the brain to form a three-dimensional f-image, and it becomes impossible to perceive the image clearly as a nuisance. The range of deviation within which a three-dimensional image can be formed differs from person to person, and it is said that the limit can be expanded through training, but it is generally said to be around 0.15H.

Clear viewing distance l1 for a 20-inch TV receiver (H=300m)
11! When observed from a distance of (D, = 2,100 ■), the range of deviation is 45 words. That is, Fig. 1(a)
As shown in , the right g11 is shifted to the left, and the left g11 is shifted to the right by 225 cm. When observing this, the solid 13 is about 7ii in front of the screen 10 (7) f = 2.739H = 8
It concludes at the position of 20■.

In addition, as shown in Figure 1 (b), #14 with no deviation is placed in the position of the screen, so all the deviations of the reproduced image that should be placed in front of the screen are caused by the old damage being shifted to the left, and the left image being placed on the right side. It's off. In other words, if you shift the old wound to the left and the left image to the right, the 3D lens will appear to jump out from the screen.On the other hand, if you shift the old wound to the right and the left image to the left, the 3D image will appear as if it were 4. An image appears behind the surface. At this time, since the range of deviation between old scratch 5 and old scratch 6 is 45 square meters, the limit of device position 4 is rearward of the screen b=12,6H=3,780 square.

In other words, the available space for the device is a straight line that includes three sides: an oblong furnace-shaped surface of approximately 0.39 HX O, 52 H at a position of 2.74 H in front of the screen, and an oblong rectangular surface of approximately Z8 H x 3.73 H at a position of 12.6 HO behind the screen. This is the range surrounded by .

According to this consideration, the space that can be visualized with a 20-inch TV set at 300 meters is 12,000 meters long. ．． At the position of 74H = 820cm, width 0.52H = 208cm, height 0.39" H =
117+w rectangular surface and 1 screen (400+wX300■)
And, at the rear of the screen IZ6H=786■, the width is 3.73H=1,120. aii Sa2.8H=840m
It is a space containing a surface of size . II, which protrudes from this space, is difficult to bear the weight of the axle and appears blurry. Therefore, in addition to the space in front of the screen that was possible with conventional 3D TV recycling bonds, a space approximately 4.6 times larger than the front is available behind the screen, making it possible to create a 3D image. The depth is much deeper than before, allowing for a wider axle load.

This principle serves as a guideline for how to construct an original image for stereoscopic viewing of images such as computer graphics and animation. Hereinafter, the principle of the present invention will be explained using examples that are easy to understand.
□・1＼ This will be explained based on Figure 2. \ Dual cameras 11 and 12 were prepared and set on a 3D camera mount 13. The original axes of the right camera 11 and the left camera 12 of this mount 13fl can be rotated by the same angles θ and 10 in opposite directions. θ=O
At this time, the optical axes of the two cameras are both parallel to the center line of the pedestal 13, and are looking at the subject 10 located at infinity. As the subject 10 approaches the camera, the magnitude of θ increases. At this time, since the subject is placed on the intersection of the optical axes of the two cameras, the subject will not move off the screen even when zoomed in. here,
The zooming mechanisms of the two cameras are linked, and both 0! The sizes of the two are balanced so that they are both the same size. The photographic signals from the left and right cameras are input to a monitor television 15 through a mixer 4. Since the mixer switches the camera signals using the field synchronization signal, the files of the left and right cameras are alternately displayed on the monitor television 5. Since the right camera (116) and the left camera are shifted by 7μ, it looks like a double copy when viewed with the naked eye.Stereoscopic glasses with a built-in electro-optical shutter 8
, and synchronized with the field signal in a mode in which the shutter opening and closing operations are alternately reversed.
Operate the right camera so that it cannot be seen by the right eye. When I look at the screen with these 3D glasses on both eyes, I see myself from the screen.

It stands out and looks three-dimensional.

In this case, the optical axis of the camera is always aligned with the center of the subject even if the subject moves so that the subject does not go out of the field of view even when zooming. In other words, the center of the image is always located on the screen, and although the shape appears three-dimensional, it does not give the impression that the subject has moved back or forth, or that the subject itself has moved away from the screen.

In order to make it appear as if the subject itself is moving back and forth away from the TV screen, it is necessary to keep the angle θ constant. If θ is kept constant, the field of view of the camera will be limited, so 0 will be made small and images will be taken from a distance. In this case, forward and backward movements, that is, changes in position, are displayed three-dimensionally, but it is difficult to display the shape three-dimensionally.

When zooming in to see a shape in 3D, there is an inconvenience that the subject may be out of the field of view.

Furthermore, when the subject moves ahead of the intersection of the optical axes of the two cameras, it appears to jump out of the screen, but when the subject moves behind the intersection, the effect of collapsing to the rear of the screen is It is much more compressed than the effect of jumping forward.

Therefore, the movement range of the subject is generally limited to the range in front of the intersection. Therefore, g1 of the right camera
The image will always be displayed on the TV screen shifted to the left of the camera on the left.

As mentioned above, when photographing a subject with two cameras, the stereoscopic viewing technique is considerably restricted due to technical issues and the convenience of the photographer.

However, 3D images using graphic displays and animations! Since the display is not subject to the above-mentioned restrictions, a three-dimensional image can be freely constructed.

The method will be described below with reference to FIG.

To compose a three-dimensional monk with animation, right compensation 6
Prepare flat original drawings 21° and 22 of two types corresponding to 7 and the right side. These two-dimensional original images 21 and 22 are inputted to the monitor television 15 through the right camera 11 and left camera 12, respectively, and old scratches 16. Lightly shown as left trench 17.

If the subject is essentially three-dimensional, such as a real object or landscape, then of course one subject is simply photographed using two cameras, but when shooting a subject that essentially has a thousand faces, such as three-dimensional animation, In this case, it is necessary to separately prepare two systems of original images, one for the right camera and the other for the left camera.

Now, if we use the same one for the original picture of the two systems, the old damage 16 and the right side 17 of the monitor television 15 will overlap and become one, making them indistinguishable. Therefore, stereoscopic glasses 18
Even if I use , I can't see it in 3D. To form a three-dimensional image, the right f# 16 and the left f# 17 must be displayed at least partially offset. Old wounds 16 are my right servant 1
If it shifts to the left side of 7, the monk synthesized by the brain will appear to protrude in front of the screen by a distance proportional to the amount of shift, and if it shifts to the right, the synthesized monk will appear to recede to the back of the screen by an amount proportional to the shift.

Here, original pictures 21 and 22 are Furuzuke 1 on the monitor TV.
You can freely select the difference between 6 and 17 as necessary. If the deviation is too large, the operation of the device becomes difficult, so as mentioned above, the deviation δ should be kept within 0.15H.

A 20-inch TV receiver (H=300■) is placed at clear viewing distance (
D, = 2,100■) when observing from a distance of J-0
, 15H in the horizontal direction is 45■ on the screen, and the three-dimensional image is placed at a position of 820■=273H in front of the screen. - 10,000, δ = -0,15H, that is, if you shift 451 to the left and right opposite to the previous left and right, the 3D gII will be 3,78 behind the screen.
0w=IZ, 6H (image is formed at D position.

If the protrusion distance based on the deviation δ is f, then = 1 & 2
δ ・・・・・・・・・・・・・・・ (1). Let b be the depth distance to the rear based on the deviation - δ = 84 δ ・・・・・・・・・・・・・・・・・・
(2) becomes. Here, the coefficient 18 that determines f for a
．． The problem is that the coefficient 84 of b with respect to δ is about 4.6 times larger than that of δ. In other words, a negative deviation of -6 causes a larger biasing effect than a positive deviation δ. In other words, if the subject is real,
If the distance from the screen is 1IIlf and bH, and the distance is the same, then f=-b.In other words, making it stick out to the front produces a better three-dimensional effect than pulling it in to the back.

On the other hand, in the case of a composite original image, a deviation is created from the beginning. Since Hb>>f for the same size δ, the negative −
The device can be installed in a wider space than using the sense of protrusion to create a sense of depth.

Here, to create a scene in which the main character stands in a landscape, the landscape was created with negative δ as the background, and the person was created with δ minus 0. As a result, the obtained three-dimensional image differs from the image shown in the image C for a given δ n 0, in that the depth of the background is extremely wide, and the image is illuminated as if it were standing on a vast area. , it was possible to display the detailed shape of a person's face in 3D. In addition, when the distant view is δ〉-α15H1 and the tip of the protruding hand is δ)0.15H, the distant view moves away toward infinity and becomes blurry.The tip of the hand extends in front of the observer's eyes and appears blurry. Ta.

The principle of the configuration method for stereoscopic television of the present invention, which has been explained in detail based on the embodiments above, is that when the subject is a real object, the viewpoint of the camera is usually set so that δ>0, as in the conventional case. However, in order to visualize a three-dimensional image by using two systems of essentially flat original images such as animations and graphic displays, the long distance S in the backward direction is expressed as δ<0, and small irregularities are expressed as δ>0. We have found that it is possible to effectively visualize a three-dimensional image by doing this, and have described a method for constructing a three-dimensional television system based on this.

[Brief explanation of drawings]

Figures 1 (a) and (b) are diagrams showing the relationship between 3D television screen features and connections, Figure 2 is a diagram explaining the 3D television shooting method, and Figure 3 is a diagram showing the two systems according to the present invention. FIG. 2 is a diagram illustrating the principle of displaying a two-dimensional image on a monitor TV and viewing a three-dimensional image using stereoscopic glasses. 1.5...furuzuke, 2,6...right servant, 3...equipment in space, 11...right camera, 12...left camera, 13...photography stand, 14...mixer, 15.
...Monitor TV. 16...Stone statue, 17...Sister, 18...Stereoscopic glasses, ¥J 1 Figure (Good) (b) Foil 2n 1jJ3 Figure

Claims (1)

[Claims] 1. The furuzuke and the right man for composing a three-dimensional image are displayed alternately in field order on a receiver device, and a pair of original shutters that open and close alternately on the left and right in synchronization with the field are used to display the furuzuke and the right man. Two systems of plane it (# for stereoscopic vision 1
In a method for forming a three-dimensional image configured as iif#,
To correct the misalignment between the furusuke and the right side on the screen of the receiving machine, in order to bring the stereoscopic first side to the front of the screen, the first side of the stereoscopic view must be moved to the left of the right side, and the third side of the stereoscopic view must be connected to the rear of the screen. In order to make me, I shift the right side to the right side of my right side to form a reborn monk, and if the height of the screen of the Jurin machine is H, then the line connecting the screen and the observer is Approximately 2.7H forward one screen along
A solid i1 ( How to configure *.