CA2121054C - Apparatus for stereoscopically viewing imagery of scenes and objects - Google Patents

Abstract

Various types of apparatus which embody this invention provide means for stereoscopically viewing imagery of scenes or objects, without requiring that the viewers wear special glasses or other encumbrances. A
stereoscopic effect is achieved by forming two independent images of a scene or object, one for the left eye and one for the right eye of one or more viewers. The various embodiments have several major elements in common. The first element is an array of lenses, all having the same focal length. These lenses are positioned on a surface to form an array, which is preferably regular and closely-packed and in which the centres of lenses are equidistant from each other. The lens centres lay on a surface which is planar for some, curved for others, and multifaceted for still other embodiments. For preferred embodiments of this invention, the lateral shapes of the lenses range from being circular to being those of equilateral triangles, squares, or regular hexagons.

A second major element is an array of stereoscopically-related, displaced and overlapped segments of recorded or displayed imagery of scenes or objects. For some embodiments the media of display can be photographic prints or transparencies. For others the input imagery can be generated by computer, or by other electronic means, and displayed on some type of electronic display media. In particular, these displays can be of TV imagery. This array of input image segments is positioned on a surface in front of, and parallel to, the surface on which the arrays of lenses is mounted, and is separated from this surface by a distance less than the focal length of the lenses. For a particular embodiment, the displayed input image segments preferably have the same, or substantially the same, shapes and sizes as the lenses. Each displayed image segment is aligned with a lens behind it, and is a portion of the image of an entire scene or object. The portion of the entire image displayed on each segment is displaced laterally from those portions displayed on adjacent image segments by a distance which is dependent on the separation of the lens centres, the magnification of the lenses, the curvature of the surface on which the displayed image segments are positioned, and the relative angular orientation of adjacent facets of a multifaceted surface. For some embodiments, additional arrays of displayed input image segments are positioned on other surfaces, which are also parallel to the surface on which the array of lenses are positioned. The image segments in these additional arrays are formed from imagery of objects which were at different distances from the recording location. With careful positioning, virtual images of the displayed input image segments from a particular surface will be closely-matched together, so that an essentially seamless composite output image will be reconstructed on each of one or more surfaces in front of those on which the input image segments are located. Because they are located on surfaces at different distances from the lenses, these various images will exhibit parallax. The images can be viewed by one or more persons when looking through the array of lenses, as though through a window. The illusion is created that three dimensional imagery is being viewed. This is a major novel feature of this invention.

Embodiments of the invention in which the centres of the lenses are tangent to and touch a surface which is planar or concave towards the viewers are suitable for viewing; panoramic scenes or large objects over a wide range of azimuthal and elevation viewing angles. Other embodiments, in which the surface touched by the centres of the lenses is convex towards the locations of viewers, are suitable for stereoscopically viewing images of an object as seen with different perspectives from various positions around the viewing surface formed by the array of lenses.

Description

.,, BACKGROUND OF THE INVENTION
The present invention relates to apparatus for stereoscopic viewing imagery of scenes or objects. For over one hundred years, a wide range of different types of apparatus have been developed for this purpose. These apparatus have provided means for viewing one image of a scene or object with the left eye and an independent, stereoscopically-related, image for viewing with the right eye of an observer.
Beginning with the parlor stereoscope in the last century, some of these apparatus have enjoyed a certain level of acceptance in the past, at the home consumer or public entertainment level. But few, if any, have retained any significant consumer or public appeal over time.
Various reasons have been put forward to account for the failure of previous stereoscopic viewing apparatus to achieve wide scale, long term acceptance. For some apparatus, the need to provide two independent images has required that viewers wear special glasses, fitted with red-green or polarizing lenses. Encumbrances such as these have often been judged to be awkward or irritating to wear. More recently, other types of apparatus, such as electronically-switched left-right eye liquid crystal lenses, have been developed for stereoscopic viewing of computer-generated and TV imagery. These approaches have often been judged to be too complex and /or costly for all but special stereoscopic viewing applications.
Early in this century, Gabriel l_ippmann invented integral photography. This was based on the use of arrays of extremely small lenses to form three dimensional imagery over relatively wide viewing angles. Integral photography apparatus eliminated the need for wearing special glasses. But most consumers have not been satisfied with the image quality which could be achieved with this type of apparatus. Other apparatus for stereoscopic viewing of wide angle panoramic scenery on curved surfaces have used arrays of photographic prints of anamorphoscopically-distorted imagery as input imagery displays. Arrays of anamorphoscopic lenses were then used to restore the output imagery to an undistorted final form for viewing. That is, achievement of acceptable quality of imagery required use of complex viewing apparatus.
In summary, the failure of mast previous stereoscopic viewing apparatus page 3 to win wide scale acceptance can be attributed to their complexity, lack of satisfactory image quality, or lack of convenience to use.
Accordingly, it is an object of this invention to provide a means of stereoscopic viewing of imagery which is capable of creating the illusion that three dimensional imagery is being viewed.
Secondly, it is an object of this invention to provide apparatus for stereoscopic viewing which is not complex , exhibits high image quality, and is convenient to use.
Thirdly, it is an object of this invention to provide a means for stereoscopic viewing which does not require the viewer to wear encumbrances such as special glasses or helmets, etc.
Fourthly, it is an abject of this invention to provide a means for stereoscopic viewing of relatively large exteruded scenes or objects, by a number of persons simultaneously.
Fifthly, it is an object of this invention to provide a means for stereoscopic viewing of TV imagery, which does not require any change in technical standards for TV broadcasting or video programming storage.systems.
Sixthly, it is an object of this invention to provide a means to view imagery of scenes or large objects over a wide range of azimuthal and elevation viewing angles.
Seventhly, it is an object of this invention to provide a means to view imagery of objects aver a range of viewing positions around the imagery.
Additional objects and advantages of the invention will be set forth in part in the description which follows and in part will be obvious from the description, or may be learned by practice of the invention. The objects and advantages of the invention may be realized and obtained by means of the instrumentalities and combinations particularly pointed out in the appended claims.
SUNwiARY
To achieve the foregoing objects, and in accordance with the purposes of the invention as embodied and broadly described herein, a means for stereoscopic viewing 'is provided which does not require viewers to wear polarizing glasses or other physical encumbrances. It is a significant and novel characteristic of this invention that the stereoscopic effect is achieved through the use of means to form two page 4 independent, but stereoscopically-related, images of a scene or object, one image which is viewed by the left eye, and one by the right eye of each of one ar more viewers. The apparatus to achieve this consists of a number of key elements, each of these being mounted on planar, curved or multifaceted surfaces which are parallel to each other. On a first surface is positioned an array of lenses, which can be simple single element lenses, preferably having the same focal length. For same embodiments, the lateral shape of these lenses are circular. For others, these shapes are those of equilateral triangles, squares, ar of regular hexagons, etc., which can be packed together to form very closely-packed arrays. For yet others, the shapes can be those of pentagons,isosceles triangles, or that of other polygons, which may not be closely-packed together on planar or curved surfaces. The maximum lateral dimensions of these lenses can range from fractions of a cm. in some embodiments to many tens of cms.
in others. In particular, their diameters are significantly larger than those which have been used for integral photography, which are required to be very small fractions of a cm. These lenses are positioned so as to form an array,preferably with centres equidistant from each other. For some embodiments, the lens centres fie on a planar surface, for others they touch a curved surface, while for still others they lie on each of a number of facets of a multifaceted surface. For many embodiments, it is preferable for lenses to be positioned so as to form a very closely-packed regular array, which for some embodiments can be achieved when edges of adjacent lenses coincide.
A second element is an array of stereoscopically-related, displaced and overlapped segments of displayed imagery.of scenes or objects.
Photographic or electronic media can be used for input image display purposes. In particular for some embodiments , displays of frames of current conventional TV, or HDTV (High Definition TV) imagery can be used. The array of input displayed image segments is positioned on a surface in front of, and parallel to, the surface on which the array of lenses is located. This surface is separated from that on which the lenses are positioned by a distance less than the focal length of the lenses. For many embodiments, the input image segments preferably have the same lateral shapes and sizes as those of the lenses of those embodiments. Each output magnified image segment forms a portion of the complete composite output virtual image to be viewed. The portion of the image displayed on each segment is displaced laterally from those portions of the entire image displayed on adjacent image segments. The page 5 chosen displacement in image between adjacent segments is dependent on the magnification of the lenses, the distance separation of the lens centres, the curvature of the surface on which those particular image segments are located and, in the case of multifaceted surfaces, on the the relative angular orientation of adjacent facets. Each input image segment is aligned in front of a lens, which is located on the lens surface behind it. As a result, on each of the output image surfaces an complete magnified virtual image is reconstructed which is composed of magnified segments of the input image segments on each particular input image surface. These output images are formed on surtaces which lie ahead of those on which the image segments are displayed.
It is a significant and novel characteristic of all embodiments of this invention that the left eye and the right eye of each viewer will each see independent composite virtual images. For various embodiments, the perspective presented by these two independent images to a viewer will shift with change in viewing position. For other embodiments the perspective will not change. Nevertheless, for all embodiments an illusion is created that the image being viewed is indeed three dimensional.
A further significant and novel characteristic of this invention is that stereoscopic TV viewing apparatus can be built which in the limit requires the transmission and display of a sequence of frames of only a single channel of standard or high definition two dimensional TV
imagery. Further, other apparatus can be built which requires the transmission and display of only a few channels of TV imagery.
Embodiments of this invention such as these, which require only a limited number of TV channels to achieve various levels of stereoscopic viewing capabilities, do not have the complexities and costs of those embodiments of this invention which require the capability to provide multiple channels of TV.
For some embodiments of the inventionadditionalarrays of stereoscopically-related, segmentsof other displaced, and overlapped imageryare positioned on other surfaces in f that which the front o on lenses or fartheraway from, are located.
These arrays may be closer to, the array of images segments from the on whichthe lens first surface array located. The positioning of these chosen achieve is arrays is to the desiredrelative positioning of the surfaces h the on whic output magnified page 6 a virtual images are formed. For example, input imagery on the surface which is closest to that on which the lenses are positioned could be of foreground objects, and so on for middle ground and background objects or scenery. As a result, output composite virtual images are formed on each of a number of display surfaces, and these images will exhibit parallax with respect to each other. They can be seen by viewers from positions behind the array of lenses, when looking through the array of these lenses as though through a window. This capability to form images on surfaces at a number of different distances from viewers which exhibit parallax constitutes a third significant and novel characteristic of this invention.
A fourth significant and novel characteristic of this invention is realized in many embodiments when the lateral shapes of lenses in the lens array, and the image segments of each of the different input image segment arrays, are chosen such that very closely-packed arrays can be formed. As a result, adjacent magnified image segments of the output composite image can be matched so closely together at their boundaries that they will form essentially seamless composite virtual images of scenes or objects on each of one or more particular output image planes.
Further, these output images can be viewed through an array of lenses which are arranged sufficiently close together to form an essentially seamless, window-like surface.
A fifth significant and novel characteristic of other embodiments of this invention is achieved by recording imagery of an object over a wide range of azimuthal and elevation angles, and then configuring embodiments of this invention such that images of the object can be subsequently reconstructed and viewed over a range of azimuthal viewing angles of up to 360 degrees.and over a wide range of elevation angles.
BRIEF DESCRIPTION OF THE DRAWINGS
The accompanying drawings, which are incorporated in, and constitute a part of the specification, illustrate particular embodiments of the invention and, together with the general description of the invention given above and the detailed description of the preferred embodiments given below, serve to explain the principles of the invention.
FIGURE 1 is a diagrammatic perspective view of one approach for locating a camera to record imagery of static three-dimensional scenes or objects at various positions on a plane over a range of page 7 a azimuthaf and elevation angles. For some embodiments of this invention recording takes place in one direction from only one position.
FIGURE 2 is a diagrammatic view of a multiple lens photographic camera for simultaneously recording imagery of scenes or objects from various positions, for subsequent stereoscopic viewing by apparatus according to this invention.
FIGURE 3 is a diagrammatic view of an array of small TV
cameras which can be used for simultaneously recording imagery of scenes or objects from various positions, for subsequent viewing by TV
stereoscopic viewing apparatus according to this invention.
FIGURE 4 is a diagrammatic perspective view of an embodiment of this invention for stereoscopic viewing of imagery which has been recorded by various types of photographic apparatus, such as those illustrated in FIGURES 1 and 2. For this embodiment, the array of lenses and the array of image segments are each positioned on planar vertical surfaces which are parallel to each other.
FIGURE 5 is a central horizontal cross-sectional view of the embodiment of this invention illustrated in FIGURE 4, for stereoscopically viewing imagery of three dimensional scenes or objects. This figure illustrates a means to vary the lateral positions of lenses and input image segments as may be required in some embodiments of this invention in which the lenses and input image segments are arranged in the form of regular, quasi-regular, or irregular arrays. FIGURE f is a central horizontal cross-sectional view of an embodiment of this invention for stereoscopically viewing TV imagery of three dimensional scenes or objects, as recorded by TV camera apparatus such as that illustrated in FIGURE.3 , or by some other electronic apparatus for subsequent display on electronic display media. This figure illustrates a means to vary the lateral positions of lenses and TV image display units, as may be required in some embodiments of this invention in which the lenses and input image segments are arranged in the form of regular, quasi-regular, or irregular arrays.
FIGURE 7 is a central cross sectional view of the embodiment of the invention illustrated in FIGURE 4, showing light rays from various subsegments of displayed image segments tn the left and right eyes of a viewer.
FIGURE 8 A is a diagrammatic perspective view of an embodiment of this invention which is used for stereoscopic viewing the imagery of a scene or object which has been recorded in one direction from only one position, and is displayed such that an illusion is created page 8 that the image being viewed through the plane of lenses is three dimensional. For this embodiment, the lateral shapes of the lenses and the input image segments are those of regular hexagons. FIGURE 8 B
illustrates an embodiment which is similar in many respects to that illustrated in FIGURE.BA, except that the lateral shapes of lenses and input image segments are those of equilateral triangles. FIGURE 8C
illustrates an embodiment for which the lateral shapes of lenses and input image segments are squares. This embodiment also has certain image forming characteristics which are different from those of FIGURES
8 A and 8B. In particular, this figure illustrates an embodiment by means of which different output images can be sequentially reconstructed on an output image plane. This is accomplished by the process of synchronously rotating a number of rods on which rows of associated input image segments are positioned until the segments for a desired output image are matched up together to form a complete array of segments on an input image plane in front of the array of imaging lenses.
FIGURE 9 is an elevation view of one method of dividing up the plane of an input image display of a scene or object which has been recorded with a photographic or TV camera.in only direction from one position, The recorded imagery will be subsequently viewed using embodiments of this invention such as those illustrated in FIGURES 8 A , 8 B o r 8 C. FIGURE 9 shows the input display to be divided into hexagonally-shaped subsegments, as is the case of FIGURE 8 A. These are arranged to form a closely-packed array.of image segments. The figure shows five subsegments namely, C'1', ~'2', t~', Cep' and C3 ', along a central horizontal elevation in the plane of the input image segments.
FIGURE 1 0 is an elevation view of seven hexagonally-shaped input image print segments or TV frame segments which are based on use of the particular subsegments illustrated in FIGURE 9. They are here shown as separated for illustrative purposes although, in practice, they would be arranged with edges touching to form a closely-packed array of segments. Each of these seven segments is made up of seven hexagonally-shaped subsegments and six partial subsegments. There are three subsegments positioned in a horizontal direction across each image segment, with the horizontal width of image displacements between adjacent segments being one subsegment. This arrangement is for a final image magnification of a factor of three. FIGURE 1 0 shows subsegments C1, Q and C3 , across the centre of segment 10-3; (3, C3 and G4 , across the centre of segment 10-4, and C~, G4 and (1; , across the centre of segment 1 0 - 5.
page 9 s ,:
FIGURE 1 1 is a diagrammatic perspective view of an embodiment of this invention in which there are two parallel surfaces on which image segments are positioned. Image segments of foreground objects are positioned an the surface closest to the lens array, while those for background scenery and objects. are positioned on a more distant surface. As a consequence, output virtual images are formed on each of two separated output image surfaces:
FIGURE 1 2 is a cental horizontal cross-sectional view of an embodiment of the invention which is configured as a multistage apparatus for stereoscopically viewing imagery of various types of objects. Suitable objects may range from single photographic print recordings of a scene or objects through to a series of frames of recorded or transmitted TV imagery which are displayed on a conventional TV
display unit.
FIGURE 1 3 is a diagrammatic perspective view of a multistage embodiment of this invention which is configured for stereoscopic viewing of displayed TV imagery. This embodiment requires displaying only a single channel of transmitted TV signals in order to create an illusion that the TV imagery being viewed is three dimensional.
FIGURE 1 4 is a diagrammatic perspective view of an embodiment in which the centres of the lenses and input image segments touch, and are tangent to, concentric spherical surfaces which are concave towards viewers. The output composite imagery of background scenery can be viewed on the more distant spherical output imaging surface over a range of azimuthal angles of up to 180 degrees, and elevation angles of up to 90 degrees from locations in the vicinity of the centre of the concentric spherical surfaces. In addition, the output composite imagery of foreground objects or persons can be viewed on a closer spherical surtace.
FIGURE 1 5 is a diagrammatic perspective view of an embodiment in which the surface touched by the centres of the lenses is cylindrical, and is oriented such as to be concave towards the viewers.
Output virtual imagery can be viewed from locations near the axis of the cylinder.
FIGURE 1 6 illustrates a horizontal equatorial cross-sectional view of a portion of the embodiments illustrated in FIGURES 1 4 and 1 5, for which the viewing surface which the centres of the lenses touch, and are tangent to, is concave towards the viewers.
FIGURE 1 7 is an elevation view of five hexagonally-shaped photographic print or TV image segments, which could be used as input page 10 image segments for embodiments such as that illustrated in FIGURE1 5, and for which the lenses and input image segments are mounted with their centres touching concentric cylindrical surfaces. Because lenses and input image segments are located on cylindrical surfaces which are concave towards viewers, the horizontal displacements of the subsegments of the image segments of FIGURE 1 7 are necessarily larger than those displacements of the subsegments of input image segments for planar surfaces as illustrated in FIGURE 1 0.
FIGURE 1 8 is an elevation view of a strip of output composite virtual image, as formed by imaging the input image segments illustrated in FIGURE 17.
FIGURE 1 9 A is a diagrammatic view of one arrangement for positioning a camera to record imagery of a panoramic scene in each of a small number of widely-separated azimuthal directions. FIGURE 19 B is a diagrammatic view of an embodiment of this invention which can be used for stereoscopic viewing of imagery recorded by an arrangement such as that illustrated in FIGURE 1 9 A.
FIGURES 2 0 A and 2 0 B are diagrammatic perspective views to illustrate several methods of pasitioning cameras on convex cylindrical surfaces in order to record imagery of a person or object in a number of azimuthal directions around the person or object. FIGURE 20A illustrates sequential recording of imagery at a number of positions over a range of azimuthal angles of up to 360 degrees. FIGURE 20B illustrates simultaneous.recording of imagery in a number of azimuthal directions.
FIGURE 2 1 is a diagrammatic perspective view of an embodiment of this invention for stereoscopically viewing imagery of the person recorded using methods such as those illustrated in FIGURES 2 0 A
or 20B. Output reconstructed images of the person are seen to be contained within a cylindrically-shaped surface which is concentric with that touched by the centres of the lenses.
FIGURE 22 is a horizontal cross-sectional view of the embodiment of the invention illustrated in FIGURE 2 1, which illustrates formation of output imagery on each of a number of planes, which can be viewed from a number of azimuthal directions.
FIGURE 23A is a diagrammatic view of the multifaceted lens surface for an embodiment of this invention in which an array of equilateral lenses is mounted on the surface of an icosahedron. FIGURE
23B illustrates this surtace as it would be seen if it were to be flattened out onto a plane. FIGURE 23C illustrates another multifaceted lens surface for which the lateral shapes of the lenses are those of both page 1 1 t t equilateral and isosceles triangles, as this surface would be seen if it were to be flattened out onto a plane.
FIGURE 2 4 is a diagrammatic view of.the multifaceted surface of a truncated icosahedron, which is a solid which has twenty hexagonal and twelve pentagonal facets . As illustrated in FIGURE 2 4, arrays of lenses can be positioned on some or all of these facets in some embodiments of this invention. Similarly, arrays of image segments can be positioned on the faces of one or more concentric truncated icosahedrans in order to provide a capability to view stereoscopic imagery from a large number of directions.
FIGURE 2 5 is a diagrammatic view of the well-known geodesic dome invented by Buckminster Fuller, which is based an use of arrays of equilateral triangles of various sizes to form a multifaceted surface.
Arrays of lenses and image segments shaped as equilateral triangles can each be positioned on these facets to form many embodiments of this invention if it is required to provide a capability to view stereoscopic imagery over essentially ali directions.
FIGURE 2 f is a diagrammatic view of a recently invented shape known as a geotangent, or ellipsoidal, dome. This nonspherically-shaped surface may be chosen as a surface on which to position tenses and input image segments for various embodiments of this invention.
FIGURE 2 7 A is an elevation view of an embodiment of this invention in which the lateral shapes of lenses, and those of the associated input image segments, are regular and nearly-regular hexagons which touch a spherical surface. FIGURE ~7B is a horizontal cross-sectional view of the embodiment of FIGURE ~ T A.
DETAILED DESCRIPTION
FIG.1 illustrates a method for photographically-recording images of three dimensional static scenes or objects, for subsequent stereoscopic viewing using various embodiments of this invention. A
camera 1 is positioned on vertical plane 2 at location 3-1 and directed towards the scene or ob;ject~ ts> be recorded. For some embodiments of this invention , such as those illustrated in FIC~S.4, 5 and 6, imagery is recorded at each of a number of positions ~-14 3-2, 3-3, etc, which may be separated from each other by the interocular distance of viewers eyes of about 6.4 cm. After completion of the recording at positions along one row over the distance necessary to cover the desired azimuthal angular viewing range, the camera is returned to the beginning of this row. It is then displaced downwards and sideways. 'The downward displacement is page 1 2 preferably between about 5.6 and 6.4 cm, and the horizontal sideways displacement is offset by a distance of about 3.2 cm. 'Then recording is recommenced along the next row, with same horizontal separation of positions of 6.4 cm. This procedure is repeated until an array of displaced and overlapped images of the scene or ohject has been recorded over the range of distances required to cover the desired vertical and horizontal angular viewing range. The downward displacement of 5.6 cm. is chosen to result in recording at positions which are equidistant from adjacent recording positions. Except at edges, there will be six equidistant adjacent recording positions. This pattern of recording; is then matched to a preferred geometry of the arrays of imaging lenses and input image segments in various embodiments of this invention for stereoscopic viewing.
For some embodiments of this invention, such as those illustrated in FIGS.8A, 8B, 8C, 12, 13, 14, 15, 17, 18 and 2 3 , recording of a scene or object is required in only one azimuthal direction from only one position, such as location 3 - 1 of FIC~.1 . For other embodiments, recording may take place at a number of irregularity-spaced positions on plane 2.
For still other embodiments, recording rnay take place at only a small number of positions. F1G.2 is a diagrammatic: perspective view of a multiple lens photographic camera for recording imagery of three dimensional static or moving objects at a small number of different positions. Seven objective lenses, 4 -1 to 4 - 7 , of the camera lens array are illustrated in this figure, although for some embodiments a larger number of objective lenses may be required in order to cover the angular field of view required. A side of the camera body, e.g., 5 - 1, is shown cutaway, to exhibit one of a number of hollow light filter tubes, e.g., 6-1.
Each of these tubular filters, e.g., 6 -1, is mounted between a lens, a g, 4 -1, and the particular area, e.g.. 7 -1 , of the emulsion surface, onto which the image formed by this lens is focussed. The inside of these light tubes . are lined with light absorbing material in order to cutoff the extraneous light which could otherwise enter the tube from peripheral angular directions. Each camera subsystem consists of an objective lens, its associated light filter tube and its associated segment of the emulsion. A
number of these camera subsystems are mounted together in a closely-packed array configuration as illustrated in FIG.2. This arrangement of subsystems, together with a shutter for simultaneously exposing the film in all subsystems, forms a photographic apparatus which is suitable for simultaneously recording imagery of persons, scenes or objects over a page 13 ' CA 02121054 2002-11-30 a relatively small angular range as may be required for various embodiments of this invention.
FIG.3 is an elevation view of a camera arrangement for recording a scene or object for subsequent stereoscopic TV viewing, using embodiments of this invention such as that illustrated in FIC~.6 , A number of TV cameras, 8 - 1 to 8 - 7, with the centres of their objective lenses separated laterally by a distance which is preferably about that of the interocular distance, are shown to be clustered together to form a regular and relatively closely- packed camera array.
FIG.4 is a diagrammatic perspective view of one embodiment of this invention for stereoscopic viewing of an image of a scene or object which may have been recorded at a number of regularily-spaced positions using apparatus such as that described above and illustrated in FIGS.1, 2 or 3. Displays based on the use of photographic prints and transparencies, or other media, may be employed in various embodiments of this invention. In particular, various means for displaying imagery such as TV
cathode ray tubes, liquid crystal media, or other electronic displays are suitable for this and other embodiments of the invention. Photographic prints are shown in FIG.4 to illustrate this embodiment. They are mounted on vertical plane 9 of the stereoscopic image viewing apparatus.
Prints 10-1, 10-2, 10-3 and 1 4 - 4, etc, preferably have about the same lateral height and width as the diameter of imaging lenses 11-l, 11-2, 11-3 and 1 1 - 4, etc, on plane 1 x . The diameters of these lenses can range from a few tenths of a cm. up to tens of cm. That is, these lenses have much larger diameters than those used in stereoscopic viewing systems based on integral photography. The use of such relatively large diameter lenses constitutes a major difference between all embodiments of this invention and those systems based on use of the very small lenses of integral photography. The lenses used in various embodiments of this invention can be standard types of single or multiple element lenses which are constructed of glass or plastic. As one example of suitable lenses, single element Fresnel-type lenses of standard plastic construction have been used successfully in various embodiments of this invention. The enlarged virtual image of the scene and objects illustrated in FIG.1 will then be seen on plane 1 3 by viewers 1 4 from behind plane 1 2 on which the lenses are positioned.
FIGS is a horizontal cross-sectional view of various embodiments of this invention, as exemplified by that illustrated in FIG.4.
page 14 For these embodiments, photographic prints or transparencies 10-1, 10-

2 and 1 0 - 3, etc, are mounted on plane 9 in either a regular or an irregular array configuration. As shown in F'1C3.5 these prints or transparencies preferably have about the same lateral sizes as the diameters of the imaging lenses 11-1, 11-2, 11-3, etc., which are positioned on plane 12. When one of these embodiments of the invention is used to view the output image of a deep three dimensional scene or object, the lenses may be arranged to form a relatively loosely-packed array. Furthermore, means 15-1, 1S-2 and 15-3, are provided for each of these lenses to be moved laterally with respect to each other and hence to vary the degree of alignment, or relative lateral positioning, of each lens with respect to the particular print immediately in front of it.
A print and the lens positioned immediately behind, and in alignment with it, form a subsystem of the embodiment of this invention.
A number of these subsystems are arranged together to form a complete embodiment of this stereoscopic viewing apparatus. Means 1 6 are provided to illuminate an array of prints from immediately behind the array or means 1 7 are provided to illuminate an array of transparencies from in front of the array. In some circumstances, natural lighting may be preferable fox illumination purposes.
For many embodiments of this invention, and as illustrated in FIGS.4 and 5 , prints are positioned on a plane 9 parallel to plane 12 o n which the lenses are positioned, and at a distance from the plane of the lenses which is less than the focal length of the lenses. By well-known laws of optics, the distance separation between lenses and prints can be chosen to result in a desired magnification of the output image and the positioning of the plane on which an output virtual image is formed.
Further, the lateral alignment of each particular lens with respect to the print in front of, and associated with it on the one hand, and the magnification on the other, determine the lateral position of the output image of that particular magnified image segment. The particular displacement of the image segment recorded in each print, relative to the displacement of the image segments recorded in adjacent prints, can be laterally adjusted. Under particular conditions to be described below, the boundaries of adjacent magnified segments of an image can be arranged to be well-matched to each other so that the output composite reconstructed image will be essentially seamless. This image will exhibit parallax over a range of viewing angles which correspond to that of the angular recording.
page 15 o J

It can be viewed by one or more persons 14 from positions behind plane 1 2.
FIG.6 is a central horizontal cross-sectional view of an embodiment of this invention for viewing itraagery recorded using apparatus such as that illustrated in FIG.3. A regular or an irregular array of TV monitors,of which three 18-1, 18-2 and 18 - 3, are illustrated, is used to display recorded frames of TV imagery. This embodiment for stereoscopically viewing frames of TV imagery is the same in other essential respects as those illustrated in IpIGS.4 and 5 for viewing photographic imagery.
Consider now the condition which must be met to achieve stereoscopic composite imagery which is essentially seamless. This is that the boundaries of each of the component magnified segments of the output image must be seen to be closely-matched to the boundaries of adjacent magnified segments of the image, when viewed through the array of lenses. This condition is relatively easy to meet in those circumstances when imagery is recorded of scenes or objects which exhibit little or no variation in depth. It is also easy to meet when imagery of deep scenery or objects is recorded over a relatively narrow angular range. In the limit and as described below for embodiments illustrated in FIGS.BA, 8B, 8C, 9 and 10, it can be readily met if recordings of a scene or object are made in only one direction from one position.
On the other hand, this condition may be more difficult to achieve in circumstances when imagery of relatively deep three dimensional scenery or objects is recorded a~ad subsequently viewed over a wide range of angles. In such circumstances some parts of the output image may exhibit areas of badly-matched image segments. This imaging difficulty arises in part because the various embodiments of the invention are based on the use of relatively large diameter imaging lenses. These lenses may subtend appreciable angles when close to objects being recorded, so that imagery recorded through adjacent lenses may exhibit noticeably different perspectives and hence be poorly-matched to each other.
One of the significant and novel characteristics of this invention is that means are provided to alleviate the problem arising from poorly-matched image segments described above. As illustrated in FIGS.S and 6 , page 16 S
facilities 1S-1, 15-2 and 1 S - 3, etc, are provided in each subsystem of these particular embodiments by means of which the lateral alignment of particular array lenses 11-1, 11-2 and 1 1- 3, etc, may be adjusted with respect to their associated image print segments that is 10-1, 10-2 andl 0 - 3, etc, respectively, as shown in FIGS , or with respect to their associated TV image display units, that is 18-1, 18-2 andl8-3, etc, respectively, as shown in FIG.6. Similarly, facilities 19-1, 19-2 and 1 9 -

3, etc, are provided in each subsystem of these particular embodiments with which to effect lateral adjustments of the subsystem image print segments as shown in in FIGS , or the TV image displays as shown in FIG.6. Such adjustments in lateral alignment of lenses with input image segments can improve the matching of boundaries of particular adjacent image segments to each other, thereby improving image quality in those particular portions of the composite image.
Consider now one person looking through the array of lenses on plane 12 of the embodiments illustrated in FIGS.4, S or b as though looking through a window, and observing the reconstructed magnified virtual image. This is illustrated in central horizontal cross-section in FIG.7 , for the case of a image magnification or a factor of two. The horizontal cross-section of the image is shown to be divided into three laterally displaced image segments 10-1, 10-2 and 10 - 3. In turn each input image segment, which was recorded from a slightly different angle than were the others, consists of four overlapping subsegments. That is to say, segment 10 - 1 recorded at one angular direction will be seen to consist of subsegments Al - 1, A2 - l, A3 - 1 and A4 - 1 when viewed in cross-section. FIG.7 shows that for the particular viewer location illustrated, the left eyel 4 - 1 is so positioned relative to the array lenses 11-l, 11-2 and 11-3 that it will see subsegments Al to A6 of the complete image, when looking, through lenses 1 1 - 1, 1 1- 2 and 1 1 - 3. On the other hand, the right eye 14-2 will see the subsegments A3 to A8 when looking through these lenses. More generally, when looking through the entire array of lenses from various distances the left eye will see a composite image which is stereoscopically-relaited to, displaced from, and independent of the image seen fly the right eye.
This is a significant feature of this and all embodiments of the invention. That is, this invention provides a novel and straightforward method to provide two independent images, one for each eye of one or more viewers. This is based on particular arrangements of image page 17 s segments and the positioning of simple lenses behind them.for the various embodiments. By such arrangements these embodiments eliminate need for polarizing glasses, the complexities and expenses of special distorting lens combinations, and are not subject to the inadequacies of integral photography image quality. These are the prices which have had to be paid far using one or another of the various other approaches to meeting the need to provide two independent images in stereoscopic image viewing systems.
FIG.B A illustrates an embodiment of' this invention which is based on a novel approach to avoid the problems arising from recording deep scenes or objects aver wide angular ranges. This approach is based on the strategy of recording such scenes or objects at one position and in one direction only, for example recording with the camera located only at position 3 - 1 in FIG.1 and recording imagery in only one direction. This approach can be successful because of a little-known characteristic of stereoscopic vision. This is that by using an appropriate type of apparatus for stereoscopic viewing, an illusion can be created that a three dimensional image is being viewed, even if the input imagery which is recorded is that of an object as seen in one direction from one position.
The embodiment illustrated in FIG.BA achieves this by displaying an array of images of the two dimensional scene or abject such that the imagery to be seen by the left eye of an observer is stereoscopically-related to,offset from, and independent of the image of the scene or object to be seen by the right eye of the observer.
The embodiment shown in a diagrammatic perspective view in FIG.BA consists of elements having essentially the same functions as those of the embodiment illustrated in FIG.4 for viewing imagery of scenes or objects which had been recorded aver a range of angles. The first element is an array of imaging lenses 20-l, 20-2 and 2 0 - 3, etc, mounted on a vertical plane 1 2. Far this embodiment, the lateral shape of each these lenses is chosen to be that of a regular hexagon, and they are arranged together in the form of a very closely-packed ~uray.
The second element is mounted an a vertical input image display plane 9. It consists of an array of input image segments 21-1, 21-2 and 21- 3 , etc, which could be photographic prints, photographic transparencies, or TV image display segments of the complete input image display. These segments are stereoscopically-related, displaced and page 1 8 r overlapped versions of each other. The array of recorded image segments is positioned in front of, and aligned with, particular lenses 2 0-1, 20-2, 2 0 - 3, respectively. The lateral shape of each of the image segments in FIG.BA is that of a regular hexagon, having the same size as that of a lens.
Taken together, a lens 2 0 - 1, e.g, and the particular input object segment 2 1 - 1 in front of and in alignment with it, form a subsystem of the complete apparatus. The plane 12 is so positioned longitudinally that the distance between it and plane. 9 in front of it is less than the focal length of the lenses. As a consequence, a magnified virtual composite image of the input object display is formed on plane 1 3 which lies ahead of plane 9.
The principal differences between the embodiments illustrated in FIG.BA and FIG. 4 arise from the differences in strategy adopted for choosing the input image segments and their relative displacements. By following this strategy major objects of this invention can be realized. First, the displacements of input image segments are chosen so that the output magnified virtual imagery segments are closely-matched together. Secondly, the shapes of lenses and input image segments are chosen so that they can be arranged to form closely-packed arrays. As a result the output composite image will appear to be essentially seamless and can be viewed by a number of viewers at the same time. Embodiments exemplified by that illustrated in FIG.8 A
achieve this by choosing regular hexagons as the lateral shape of lenses 20-1, 20-2 and 2 0 - 3 , etc, and input image segments 21-1, 21-2 a n d 21-3, etc, and arranging these lenses and input image segments into closely-packed arrays. In a similar manner, embodiments exemplified by that illustrated in FIG.8 B achieve this by choosing equilateral triangles as the lateral shape of the lenses, 22-1, 22-2 and 2 2 ~ 3, etc. and input image segments 23-1, 23-2 and 23-3, etc. Further, embodiments exemplified by that illustrated in FIG.BC achieve this by choosing squares as the lateral shape of the lenses 24-1, 24-2 a n d 2 4 - 3 , etc, and the input image segments 25-l, 25-2 and 25-3, etc and arranging these together other as shawn, FIG.BC also illustrates a means of mounting three different sets of input image segments in front of the one array of lenses 2 4 - 1, etc, in such a manner that three different reconstructed output images can be sequentially selected and and presented for viewing. As illustrated in FIG.BC, each horizontal row of image segments 2 5 - 1, etc, associated with page 1 9 one particular output image is mounted on ore of the three faces of a horizontal mounting rod 26~1, etc, which has a cross-sectional shape of an equilateral triangle. These mounting rods are positioned one above the other on a vertical place. Means are provided to rotate each of these rods about an axis 2 7 - 1, e.g, which runs through the centroid of its cross-section. In operation, the rods are rotated until the input image segments associated with the first output image on each rod are aligned with each other and lie on the vertical plane 9. The first output image is then reconstructed on plane 13. Similarly, when it is required to reconstruct another output image on plane 1 3 , all of the rods are rotated together through an angle of 120 degrees so that input image segments associated with the second complete output image will lie on the plane 9. Rotation of the rods through a further 120 degrees results in the image segments associated with a third output image becoming aligned and lying together on the plane 9. In other similar embodiments,which are not illustrated here, the rods may have thin rectangular cross-sections, thus providing a capability for displaying either, or indeed neither, of only two sets of input image segments. In still others, also not illustrated here, the rods may have square or polygon-shaped cross-sections FIG.9 is an elevation view of a method of dividing up the the plane of an image display of a scene or object, which has been photographically recorded at only one pasition, e.g., 3 - 1, and in one direction, as illustrated in FIG. 1. This recorded image will subsequently be viewed using an embodiment of this invention such as that illustrated in FIG. 8A. The input image display is divided into hexagonally-shaped subsegments, arranged in the form of a closely-packed array. FIG.9 shows five subsegments Cl ', C2 '~3 ', C4 ' and CS ' along a central horizontal elevation section of the input image display.
FIG.10 is an elevation view of an arrangement of seven hexagonally-shaped input photographic image prints or TV image frame segments 2 1 -1, 21-2, 21-3, 21~4, 21-5, 21-6 and 2 1 - 7 which has been formed by positioning the subsegments of FIG.9 together as a closely-packed array.
In FIG.1 0 these segments are shown to be separated from each other, for illustrative purposes only. In practice they would be arranged to form a closely-packed array with no distance separating them. Each of these seven segments is madt: up of seven hexagc>nally-shaped subsegments and six partial subsegments. There are three subsegments in the horizontal elevation section of each image segment, with a horizontal displacement page 2 0 between those adjacent segments which is chosen to be the width of one of the three subsegments. This is the required displacement for a final image magnification of a factor of three. For example, FIG.10 shows a central horizontal elevation section, consisting of subsegments Cl , C2 and C3 of segment 21- 3 ; CL , C3 and C4 of segment 21- 4 ; and C3, (~ and CS , of segment 2 1 - 5. FIG.1 d also shows that the subsegments C1 to C5 in the upper segments 2 1 - 1 and 2 1 - 2 are displaced vertically downwards by one third of the height of the segment. Similarly, the same subsegments in the lower segments 21 -6 and21 - 7 are displaced vertically upwards by amounts appropriate for a magnification of a factor of three.
More generally, for input image segments mounted on a planar surface, the correct displacement of images between adjacent print segments will be equal to the separation of the centre lines of the lenses, divided by the magnification of the output image.
For images recorded at only one location and in one angular direction, the above choices crf relative displacements of image subsegments in each input image segment ensure that the boundaries between the magnified segments of the output image on plane 1 3 will be closely-matched with each other. This has several favourable consequences for output image quality. First, the resulting composite image will be essentially seamless. That is, there will be no ill-matched boundaries between segments to detract from overall image quality, such as may arise when viewing imagery of deep three dimensional input objects which may have been recorded over a wide angular range.
Secondly, because there is no longer any uncertainty about the required relative displacements of adjacent input image print segments, or of TV
imagery displayed on each of the monitor units to achieve the most satisfactory output image far viewing, the arrays of final imaging lenses and input image segments do not necessarily have to be adjusted laterally, as may be the case for embodiments such as those illustrated in FIGS. 5 and 6. Instead, the arrays can be arranged to form very closely-packed arrays. For various embodiments of this invention, the design of such closely-packed arrays can be based on choice of regular hexagons as lateral shapes for the lenses and input image segments, as illustrated in FIGS. &A, 13, 24 and 27. Alternately, these shapes can be those of equilateral triangles, as illustrated in FIG. 8 B, or squares, as illustrated in FIGS.8C and 15 , Further, by using these shapes, the arrays of lenses can be constructed in the fc>rm of essentially continuous, window-like surfaces, page 2 1 ...... ,..~, ~.~,.. .... . ...... -~..-.~..~,~,-......

V

through which composite and essentially seamless stereoscopic imagery can be viewed on plane 13.
FIG.1 1 is a diagrammatic perspective view of an embodiment of this apparatus for which there are two surfaces an which input image segments are positioned and for which the input image segments have been recorded as described and illustrated in FIG.4. Input image segments 28-2, 28-3, etc., of foreground objects are positianed on vertical plane 2 9, which is positianed closer to the plane 12, on which the lenses are located than is plane 9 , an which the input image segments of the background scenery and objects are located. As a result, a magnified virtual image of the foreground objects will kre seen by viewers 1 4 to be reconstructed on plane 3 0 which lies closer to the plane 12 than does plane 1 3, on which the magnified virtual imagery of the background scenery and objects are formed. Far some embodiments, the input image segments for foreground imagery.on one plane could be positioned on the planar faces of rods as described far the embodiment illustrated in FIG.BC.
By rotating these rods, output foreground imagery formed on one output image plane can be changed without affecting the output imagery formed on other planes.
Various embodiments may have a large number of input image surfaces. For many of these embodiments it may be desirable that correct size and positional relationships exist between output magnified images on different surfaces. This can be accomplished by appropriate choice of relative sizes and longitudinal positional displacements, and hence output image magnification, for the input image segments on each surface.
FIG.12 is a central horizontal cross-sectional view of a different type of embodiment of this invention in which displays of scenes or objects are directly presented as inputs for stereoscopic viewing. Suitable inputs could range from a single image display of a scene or object as viewed from one location, through to a display of a series of frames of TV imagery, as presented on the screen of a single TV
display unit. This particular embodiment differs from those illustrated in FIGS.BA, 8B and 8C in that only one input object display is required.
That is to say, instead of preparing an array of previously recorded, displaced and overlapping input image segments, this embodiment includes one or more additional optical stages for directly forming an array of displaced and overlapping image segments of the desired object page 2 2 within the apparatus itself. In other words, it can be described as combining many of the Characteristics of image recording systems such as those illustrated in FIGS,1, 2 and 3 with embodiments of this invention such as those illustrated in FIGS. 8A,88 and 8C.
In FIG. 12 a photographic print is used as the input object used in this embodiment. The print is mounted upside down on plane 3 1. The method illustrated in FIGS.9 and 1 0 for dividing up the plane of an input image display to form a closely-packed hexagonally-shaped array is also used for this embodiment. Central cross-sections of the particular subsegments L, K, J, I, and I~ are shown on plane 3 1 of FIG.1 2. Light from all subsegments of the input object display on plane 3 1 falls on a closely-packed array of lenses on plane 32. This array constitutes the primary imaging lens array and in this embodiment it is chosen to have an output image magnification of a factor three. 'Three lenses 33-1, 33-2 and 3 3 - 3 of those in the primary imaging lens array are illustrated in FIG.12. One lens, or a system comprising more than one lens, is used to form each of the input image segments required in the array. These lenses, or lens systems, each have the same focal length and are mounted on the vertical plane 32 at a distance greater than the focal length.from the plane of the input object display 3 1. As a result, an array of upright real images is focussed an a vertical plane 34 located behind the plane 3 2 at a distance greater than the focal length of these lenses. A sheet of imaging frosted glass or plastic is positioned on plane 34 onto which this array of images.is focussed. Plane 34 then becomes the input image display plane. This embodiment is sa configured that the relative displacements of adjacent segments of the focussed images on plane 3 4 are the same as those of the adjacent displayed prints of the input object display on plane 9 of the embodiments illustrated in F1GS.8A, 8B and 8C. As a consequence, the embodiment illustrated in FIG.12 achieves tire object of providing a means for stereoscopic viewing magnified output imagery, as do the embodiments illustrated in FIGS.8A, 8B and 8C. That is, it provides a means for the reconstruction of a magnified output virtual composite image on plane 3 5 which will have a seamless appearance when the boundaries of adjacent magnified segments of image contributions are closely-matched with each other. For this embodiment these particular image segments, and only these segments, of the complete images of the input object display are imaged on plane 3 4 . Means are provided for extraneous light to be filtered out of each particular image segment on plane 34, particularly light from adjacent image segments. As illustrated page 2 3 , in FIG.1 2, this filtering is such that only light incident on the lenses on plane 3 2 over the correct range of angles will contribute to the formation of images on plane 34. The required filtering is accomplished through the use of hollow tubular light filters 36-1, 36-2 and 3 6 - 3, etc, with their inside surfaces lined with light-absorbing material. Each of.these tubular filters is positioned with a lens, e.g., 3 3 -1 ,.mounted at one end of the filter and the particular image segment display 3 7 - 1 associated with it mounted at the other, as shown in FIG.12. At the lens end, these tubes preferably have the same cross-sectional width and shape as the lateral shape of the lenses themselves. Far example, they may have a hexagonal shape of the same width. Also at the image display area end, a tubular filter preferably has the same width and shape as that of a final imaging lens, e.g., 20-1 on plane 1 2. For example, the filter may have a hexagonal shape of the same width as this lens. As.a result of this filtering, and as illustrated in FIG.12 for image segments in the central horizontal cross-section, the image segment display e.g., 3 7 - 1 exhibits only the displaced images of subsegments H', I' and J', the images of subsegments K, L and all others being filtered out through absorption by light filter 36-1 ; the image segment display 3 7 - 2 exhibits only the displaced images of subsegments I', J' and K', the images of subsegments H, L, and all others being filtered out through absorption by light filter 36-2 ; and the image segment display 37-3 exhibits only the displaced images of subsegments J', K', and L'.,the images of subsegments H, I, and all others being being filtered out through absorption by light filter 3 6 - 3.
The lateral distance separation of optical centre lines chosen for the primary image array lenses on plane 32 is that which is necessary to achieve the correct relative displacement of particular image segments focussed onto plane 3 4 , for example, the correct relative displacement of one image segment, eg., 3 7 - l; relative to that of an adjacent image segment, eg, 3 7 - 2. As a result, a magnified virtual composite image with closely-matching boundaries between the magnified image segments will be viewed as a complete and seamless image on plane 3 5 . In other words, the same object has been achieved as that fc~r the embodiments illustrated in FIGS.$A, 8B and 8C, but without the need to create the input image segments in separate image recording and image segment production apparatus.
Another embodiment of this invention as a multistage apparatus can be used for viewing of stereoscopic TV imagery. In this page 2 4 c r embodiment a sequence of individual frames of displayed TV imagery, which have been recorded from one TV camera position according to conventional practice before being transmitted or stored, becomes the input object display. 'this embodiment is illustrated in a diagrammatic perspective view in FIG.1 3. A TV set is shown positioned on its back with its image display surface 3 8 an a horizontal plane 3 9 so that the display surface is facing upwards. As illustrated in F1G.1 3, the set is oriented so that the first line of the scan 40 exhibits the capital letters A to G of which only the letter G is visible in this figure. The input to the set is single channel of standard TV signal transmission which has been .received using a broadcast receiver or a video recording apparatus. Frames of TV
imagery are displayed on the set in a conventional manner. After reflection by inclined mirror 41, light from a displayed frame falls on the closely-packed array of primary imaging lenses 33-1, 33-2, etc, of which only 33-6 is fully seen in this figure. The other elements in this embodiment exhibit the same characteristics as those of the embodiment illustrated in FIG.12. Hollow tubular light -absorbing filters 36-6 cg, extend from each of the lenses 3 3 - 6, cg, to the area of the particular image segment 3 7 - 6 , cg, on plane 34 with which it is associated. FIG.13 shows this filter cut away to exhibit the lens, 3 3 - 6. Each of the final imaging lenses, 2 0 - 6, e.g, an plane 12 are used to form a segment of the magnified composite virtual image on plane 3~. This image may be seen by one or more viewers 14 behind the piane of lenses 12.
The imaging characteristics of the second stage of this multistage embodiment for stereoscopic TV viewing are essentially the same as those of the single stage embodiments for stereoscopic viewing of static photographic imagery which are illustrated in FIG5.8A, 8B and 8C. As a consequence, the advantages o1~ stereoscopic viewing of imagery recorded in one direction at only one position and which have been described for other embodiments of this invention also apply to this multistage embodiment for viewing TV imagery. In particular, this embodiment provides a relatively straightforward means to switch from a conventional mode of TV viewing to that ilU.istrated in F1G.1 3 for stereoscopic TV
viewing. This switchover can be accomplished simply by appropriately reorienting the TV set relative to the multistage stereoscopic viewing apparatus of this embodiment to foam an arrangement such as that illustrated in FIG.1 3.
The multistage embodiment of FIG.1 3 can provide economically page 2 5 important advantages. This is because it is significantly less expensive than embodiments such as those illustrated in FIGS.BA, 8B and 8C which require an array of 'rV image display units, such as illustrated in FIG.6.
On the other hand, the upper limit on achievable image quality of any multistage apparatus, exemplified by those illustrated in FIGS.12 and 13 , will necessarily be somewhat less than that achievable with a single stage apparatus.which has the much greater transmission and display resolution capability afforded by use of many TV channels.
Variations of the embodiment illustrated in FIG.l 3 could be constructed with an output composite image size which is significantly larger than that which can be displayed on standard size TV display tubes.
Upper limits on the size of the output virtual image depend on various factors. These include the number of picture elements(pixels) per line and lines per frame of currently -available TV technology. In the future, introduction of HDTV (high definition TV )technology, for example, that based on proposed MUSE standards( 1125 lines/frame), or HD-MAC
standards( 1250 lines/frame) could, in principle, lead to adoption of larger image sizes with high image quality. A second factor is output image brightness which could fall below acceptable levels if the magnification is too great. Embodiments of this invention which use a fibre optic faceplate on a 'TV display tube can overcome viewing problems resulting from low image brightness. Finally, a number of embodiments, exemplified by that illustrated in FIG.1 3, can be positioned together to form other embodiments capable of forming stereoscopic imagery on each face of a multifaceted surface.of ehe type similar to those illustrated in FIGS.19B,22, 24, 25, 2C~, an d 27 and this multifaceted TV image surface can then be used to view imagery over a wide angular range. For such embodiments, imagery formed on each face could require transmission and reception of one channel of TV broadcasting or pragramming.
In summary, various embodiments of this invention can be constructed for stereoscopic viewing of TV imagery as exemplified by the embodiment illustrated in FIG 1 3, for which only one TV image display unit is required, or based on embodiments as exemplified by those illustrated in FIGS 6, 8A, 8B, 8C, 22, 24, 2~, etc, for which a number of TV display units are required, each displaying one channel of TV
broadcasting or programming signals. In the future, introduction of HDTV
broadcasting will provide increased opportunities for stereoscopic viewing of high quality TV imagery on wide area display units.
page 2 b FIG.14 is a diagrammatic perspective view of an embodiment of this invention which is suitable for viewing stereoscopic imagery over azimuthal angles of up to 180 degrees and elevation angles from the horizon to the zenith. Viewing of imagery over such a wide angular viewing range is possible from near central locations such as 14 - 1.
Viewing imagery over narrower viewing ranges is possible from other viewer positions such as 14 - 2 and 14 - 3. The spherical surface 4 2 which is touched by the centres of the imaging lenses, and to which the surface of the lenses are tangent, presents a concave surface towards viewers. For this embodiment, the lateral shape of each lens, e.g., 1 1 - 1, and input image segment, e.g., 10 - 1, are circular. For the embodiment illustrated in FIG.14, input image segments for background imagery are positioned on surface 4 3 so that magnified virtual background imagery will be formed on surface 4 4. In addition, input image segments for foreground imagery are positioned on surface 45, resulting in magnified foreground imagery being formed on surface 4 6.
Lenses and input image segments for embodiments which provide wide angular viewing ranges may have lateral shapes chosen to produce essentially seamless composite imagery. As described below, these lateral shapes may be hexagonal, as illustrated in FIGS 24 a n d 2 7, or triangular, as illustrated in FIGS .23A, 23 B, 23C a n d 2 ~.
FIG.15 is a diagrammatic perspective view of an embodiment in which one cylindrical surface 47 is touched by the centres of the lenses and another surface 48 is touched by the the input image segments. The surfaces of the lenses and image segments are tangent to these surfaces, respectively. The cylindrical surface 4 7 is concentric with surface 4 8 and is oriented to be concave towards viewers 14 with its axis being vertical.
For this embodiment, the lateral shapes of the. lenses 2 4 - 1, etc, and input image segments 2 5 -1 , etc, are square, although other shapes such as hexagonal may be used. The complete output magnified image is formed on cylindrical surface 4 9 . From central viewing locations, eg, 14 -1 , viewing imagery over a range of azirnuthal angles up to 180 degrees is possible. From other locations, eg, 14-2 and 1 4 - 3 , viewing imagery over somewhat narrower range of arimuthal angles is possible.
FIG.16 is a central equatorial cross-sectional view of the embodiment of this invention, which is illustrated in FIG.14. The centers page 2 7 a of lenses 1 1 - 1, 11-2 and 1 1 - 3 touch a spherical surface 42 and the surfaces of these lenses are tangent to this surface. Subsegments A,B,C
and D are exhibited in cross-section on input image segment 10 -1, subsegments C, D, E and F on input image segment 10-2, and subsegments E, F, G, and H on input image segment 1 0 - 3. The centres of these input image segments touch spherical surface 4 3. 'fhe magnified output images of these subsegments, namely ,~', B', C', D', E', F', G' and H' are reconstructed on surface 4 4 to form a complete and seamless output image which can be seen by a viewer 14 from locations near the centre of the spherical surfaces.
FIG.1 7 is an elevation view of five hexagonally-shaped input image segments 21-1, 21-2, 21-3, 21-4 and 21-5 which may be displays of photographic prints, photographic transparencies, or TV imagery. The segments illustrated in FIG.1 7 are shown to be separated from one another for illustrative purposes only. In practice they would be arranged to form a closely-packed array with no distance separating them. Four subsegments can be seen in the central section of each these segments.
These may be positioned in an embodiment such as that illustrated in FIG
15, in which the arrays of lenses and input image segments are mounted with their centres touching concentric cylindrical surfaces. The cylindrical surfaces on which the lenses and input image segments are mounted in this embodiment have concave curvatures towards viewers. Because of this, the required horizontal displacements of the image subsegments shown in FIG.17 are larger than those of the image segments illustrated in FIG. 10, for otherwise similar embodiments with planar surfaces. But the vertical displacements are the same for both types of embodiments. for the same image magnification.
FIG.18 illustrates an elevation view of a strip of the output composite image,showing the magnified overlying images of segments of FIG.1 7, namely.2 1-1', 21-2', 21-3', 21-4' and 21-5', The output image to be seen to be an essentially seamless composite image of the magnified subsegments, namely C1', C~', C3', C4', C5', C6', C7' and C8'.
Various embodiments, exemplified by those illustrated in FIGS.1 4 and 1S. are suitable far viewing imagery of scenery over a wide azimuthal angular range. These embodiments provide output imagery which is essentially continuous across a wide field of view. Other embodiments, exemplified by that illustrated in F1G.19B, are also suitable for viewing page 2 8 J
imagery wide azimuthal angular range. A distinguishing feature of an embodiment such as that illustrated in F1G.19B is that the composite output imagery is made up of component imagery recorded at only a small number of azimuthal directions within a wide field of view. FIG.1 9 A
illustrates a method of recording imagery in each of a small number of azimuthal directions, which then may be processed and and positioned in an embodiment such as that illustrated in FIC:~. 19B. Imagery of the scene is recorded by a camera 1 which is positioned sequentially on each of a small number of vertical planes 2-1, 2-2, 2-3, 2-4 and 2-S which are normal to the azimuthal angle recording directions. The recording and subsequent input image segment preparation procedures required to obtain input image segments for each of these azimuthal recording directions are then essentially the same as those followed for recording and subsequent processing at only one azimuthal angle, such as those required for embodiments illustrated in FIGS.8A, $B and 8C. That is to say, each recorded image is processed to obtain an array of stereoscopically-related, displaced and overlapping image segments as illustrated in FIGS.9 and 10. Subsequently, and as illustrated in FIG. 19B, one array of square-shaped imaging lenses 24-1, 24-2, 24-3, etc, is mounted on one planar facet, e.g., 12 - 3 , of the multifaceted surface of lenses 1 2. An array of input image segments 2S-1, 2S-2, 2S-3, etc, is mounted on one planar facet 9 - 3 of the multifaceted surface of input image segment 9. These input image segments are positioned in front of, and in alignment with, particular lenses 24-l, 24-2, 24-3, etc, respectively. Output magnified imagery reconstructed from input image segments of this planar facet is then formed on output image surface, eg, 13 - 3 . Similarly, arrays of lenses are mounted on other planar facets 1 2 -1, 12-2, 12-4 and l 2 - S and arrays of input image segments are mounted on other planar facets 9-1, 9-2, 9-4, and 9-S. This results in reconstruction of output magnified images on planes 13-1, 13-2, 13-3, 13-4 and 13-S. As a result, a complete image will be reconstructed which is itself a composite of the five component output composite images. For some applications, embodiments exemplified by that illustrated in FIG.19B may be preferred over those illustrated in FIGS.14 and 1 S. For such applications recordings and image reconstruction made at only a few azirnuthal directions may suffice to obtain a composite output image which exhibits acceptable quality. For other applications, recordings and image reconstruction made at a larger number of azimuthal (and in some casr;s elevation ) angles may be necessary in order to obtain a composite outlrut image which exhibits page 2 9 i acceptable quality. For these applications embodiments such as those illustrated in FIGS. 14 and 1 5 may be preferable.
FIGS.20A and 20B are diagrammatic perspective views of camera positions on a convex cylindrical surface 5 0 for recording imagery of a person 5 1 or object, over a wide range of azimuthal angles centred on the person or abject. 'The person or object is positioned within a cylindrical surface 52 which is concentric with the cylindrical camera recording surface. FIG.20A illustrates camera locations of a camera 1 for sequentially recording views of the person or object from a number of positions over a range of azimuthal angles of up to 360 degrees. FIG 2 0 B
illustrates a method for simultaneously recording imagery of the person 1 at a number of aaimuthal angular positions 53-1, 53-2, etc, on the cylindrical camera recording surface 50.
FIG.21 is a diagrammatic perspective view of an embodiment of this invention for stereoscopically viewing imagery 54 of the person 5 1 which was recorded over a range of azimuthal angles by methods such as those illustrated in FIGS.2 0 A or 2 0 B. Inspection of FIG.21 shows that the output virtual image 54 is contained within a cylindrically-shaped surface 55 which is concentric with the cylindrical surface 56 on which the lenses 2 0 - 1, ete, are positioned. It is also concentric with the cylindrical surface 5 7 on which the input image segments 21- I , etc, are positioned. The output image 5 4 can be seen by viewers 14 from various perspectives over an omnidirectional range of viewing positions, which is the same as that over which recording took place.
FIG 22 illustrates a portion of the horizontal cross-sectional view of the embodiment of the invention illustrated in FIG.21. The centres of lenses 20-1, 20-2 and 20-3 touch the cylindrical convex surface 56. Similarly, the centres of input image segments 21-1, 21-2 and 2 1 - 3 touch the cylindrical surface 5 7. FIG.22 shows that output images are formed on planes parallel to thaw on which the input images are positioned. That is to say. when looking through a lens 1 1-1, eg, at input image segment 2 1 - r, on which subsegrnents A-1 through to L-1 are located, a viewer 1 4 - 1 will see magnified images of these subsegments on plane 5 8 - 1. Similarly, when looking through lens 20-3 at input image segment 21-3 on which subsegments C-3 through to N-3 are located, a viewer 1 4 - 3 will see magnified images of these subsegments on plane 5 $ - 3. That is to say, as the point of view of a viewer changes from one page 3 0 position to another around the cylindrical surface on which the lenses are positioned, the perspective of an image will change accordingly. In most essential respects, this is the same change in perspective as that which would be seen by a viewer looking through the cylindrical surface S 6 of lenses, if subsegments A through to N were to be positioned on the three dimensional cylindrical surface 5 5. FIG 2 2 also shows light shields 5 9 -1,59-2, 59-3, etc, which are positioned at the edges of the input image segments 21-1, 21-2 and 21-3, etc. These shields cut off viewing of distorted output imagery which would otherwise be seen when viewing imagery through the surface of lenses at oblique angles.
FIGS.2 3 A and 23B illustrate surfaces of various embodiments of this invention, for which the lateral shapes of lenses are equilateral triangles.
Arrays of these lenses are positioned on planar facets of the surface of an icosahedron. FIGS.23A and 2 3 B illustrate an embodiment in which nine lens 22-1, 22-2, 22-3, etc, , each with the lateral shape of an equilateral triangle, are positioned on each of the twenty facets of this surface. FIG.
2 3 A is a diagrammatic perspective view of this surface, and FIG.23B is a view of the same surface as it would appear of it were to be flattened out onto a plane. The input image segments for this embodiment are positioned on the surface. of another icosahedron which is concentric with that on which the lenses are positioned. Various embodiments, exemplified by that illustrated in FIG.23A,can be used by a viewer 1 4 who, when looking looking through the viewing surface formed by the array of lenses, will see imagery over a wide range of perspectives from positians around, above and below this surface, For these embodiments the array of lenses is mounted on facets on the outside surface of an icosahedran so that a convex surface is presented to viewers. Because the twenty facets of this surface completely cover° a sphere, these embodiments provide a capability to view imagery aver a greater range of elevation angles than may be possible with embodiments exemplified by that illustrated in FIG.21. Other embodiments can be used to view imagery of panoramic scenes cor objects over a wide range of azimuthal and elevation angles. For these embodiments the array of lenses are mounted on the inside surface of an icasehedron so that a concave surface is presented to viewers. Because the twenty facets of this surface cover a sphere, these embodiments provide a capability to view imagery over a range of elevation angles which is comparable to that provided by the embodiments exemplified by that illustrated irt FIG.l4.and is greater than that which can be provided by the embodiments exemplified by those page 3 1 illustrated in FIGS.15 and 19 B.
Embodiments which use the multifaceted surface of an icosahedron on which to position arrays of lenses, as exemplified by those illustrated in FIGS.2 3 A and 238, have a capability to view imagery over a wide range of angles. However this capability is nevertheless limited to only twenty relatively widely-spaced angular directions. fin the other hand, embodiments exemplified by that illustrated in FIG.23C, while similar in many respects to that of FIGS. 2 3 A and 2 3 B, have a capability to extend the number of angular viewing directions essentially without limit. This embodiment is illustrated as it would appear if it were to be flattened out onto a planar surface. Lenses 22-1, 22-2, 22-3, etc, have the lateral shapes of either equilateral or isosceles triangles. These are shown arranged in rows in which lenses 22-1,22-3, etc, in the central row,all have the lateral shape of equilateral triangles and lenses 2 2 - 2 , etc, in all other rows have lateral shapes alternating between those of equilateral and isosceles triangles. Embodiments with multifaceted surfaces which are formed with arrangements of triangles such as that illustrated in FIG.23C can be designed to have as many facets, and hence can provide as many viewing directions as may be desired.
FIG.24 is a diagrammatic perspective view of a surface composed of twenty hexagons and twelve pentagons, arranged to form the surface of a truncated icosahedron. This surface can be used in various embodiments of this invention as a multifaceted surface on which to mount lenses and image segments which have hexagonal and pentagonal shapes. FIG.2 4 illustrates how seven complete hexagonal lenses 20-1,20-2,20-3, etc, and six partial lenses can be positioned on each of the twenty hexagonally-shaped facets of this surface for various embodiments of this invention.
FIG.25is a diagrammatic view of the geodesic dome invented by Buckminster Fuller, which is based on use of a large array of equilateral triangles of several sizes to form the surface of the dome. This can be used as a surface on which to position arrays of lenses and input image segments for various embodiments of the invention in which multifaceted surfaces are required to be tangent to, and touch, a spherical surface.
Various embodiments using this type of surface may be used to provide a capability to view imagery in each of an essentially unlimited number of directions. Therefore these embodiments have a viewing capability comparable to various other types of embodiments such as those page 3 2 exemplified by the embodiment illustrated in FIG .23 C.
F 1 G . 2 6 is a diagrammatic view of a new type of dome, known as a geotangent dome, which has only recently been invented. This surface is formed of a number of hexagons and pentagons. It can be used as a surface on which to position arrays of lenses and input image segments for various embodiments of this invention in which it is required that the surfaces be tangent to, or touch, an ellipsoidal surface.
fIG. 27 A is an elevation view of an embodiment of this invention, for which the lateral shapes of lenses, and those of the associated input Image segments, are those of regular hexagons. In other respects, this embodiment is similar to that illustrated in FIG.1 4. The lenses are positioned so that the centre of each of the lenses touches a spherical surface 42. Those image segments which form the imagery of foreground objects are positioned with their centres touching the sphericat surface 4 5 which is concentric with the surface 4 2. Those image segments which form the imagery of background scenery and objects are positioned with their centres touching the spherical surface 4 3, which is also concentric with the surface 42. FiG. 27 B is a horizontal cross-sectional view of the embodiment itlustrated in FIG.27 A as viewed through section A-A.
FIG.27 B shows the concentric surfaces touched by the centres of the lenses and input image segments.

Claims (19)

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows;

1 A stereoscopic viewing apparatus providing means by which one or more viewers can view independent, stereoscopically-related, virtual images of a scene or object through a common viewing surface, in which one image can be seen by the left eye and an independent image by the right eye of each viewer, such that an illusion is created that three dimensional images of scenes or objects are being viewed; this apparatus comprising two major elements, namely, a number of lenses, each with the same focal length and each with a diameter of greater than about 0.5 cm., which are arranged beside each other on a planar, multifaceted or curved surface to form an array, and a number of stereoscopically-related, displaced, and overlapped segments of recorded images of an object or scene, which have been recorded at one or more recording positions on a planar or curved surface, each of these segments having the same, or substantially the page 3 3 same, lateral size as that of the lenses, and which are arranged beside each other to form an array on each of one or more input image display surfaces, each of which is parallel to the surface on which the lenses are positioned, and each of which is positioned a distance less than the focal length of the lenses in front of the surface on which the lenses are located, and positioned also so that each particular image segment is aligned laterally with a particular lens behind it, with the result that a particular composite magnified output virtual image is reconstructed of particular magnified displaced and overlapping image segments on each of one or more particular surfaces in front of that on which those associated particular input image segments are located.

2 A stereoscopic viewing apparatus of Claim 1, in which each lens is surrounded by adjacent lenses which are equidistant from each other such that a substantially regular array of lenses is positioned on one surface, and similiarly each of the input image segments is surrounded by adjacent imput image segments which are also equidistant from each other, with the distance separation of image segments being the same as, the distance separation of the lenses, and so that the image segments may be arranged to form substantially regular arrays on each of one or more surfaces

3 A stereoscopic viewing apparatus of Claims 1 and 2, in which means for displaying image segments onto each of one or more parallel, longitudinally-separated, image display surfaces comprises the display of an array of photographic prints, photographic transparencies or TV
imagery, on which are displayed segments of the input imagery of objects or scenes, and the longitudinal-separation of these input image display surfaces is chosen to relate directly to, or correspond to a specific location in the relative distances of these particular objects or portions of the scenery from the surface on which recording took place of the images of these objects or portions of the scenery.

4 A stereoscopic viewing apparatus of Claims 1 , 2 and 3, in which the imagery of an object or scene for subsequent stereoscopic viewing is recorded at only one recording location and in only one particular recording direction, and this recorded image is subsequently rerecorded or selected sequential segments from numerous reproductions of a single recording are used to form an array of stereoscopically-page 34 related, displaced and overlapping input image segments for subsequent display on a surface in the stereoscopic viewing apparatus of Claims 1, 2 and 3.

A stereoscopic viewing apparatus of Claims 1 , 2 and 3, to which has been added one or more stages by means of which an array of stereoscopically-related, displaced and overlapping segments of input images of a scene or object recorded from only one location and in only one particular direction, are directly focussed on to an input image display surface so that subsequently, in a second imaging stage of this apparatus, and through the use of the array of lenses according to Claims 1 ,2 and 3, a composite magnified output virtual image is formed on a surface in front of that on which the input image segments are positioned, and this apparatus also includes hollow tubes, with their insides lined with light-absorbing material to block extraneous light, and which are positioned around each of the lenses in the first: array and extend to the surface of the input image display surface.

6. Apparatus of Claims 1 , 2 , 3, 4 and 5, in which the surfaces on which the arrays of lenses and the input image displays are positioned are planar or multifaceted with planar facets, and in which one or more lenses in the array of lenses are positioned on each of these planar surfaces, or each of the planar facet surfaces, and similarly in which one or more input image segments in the array of input image are positioned on each of their own planar surfaces, or each of their own planar faceted surfaces.

7 Apparatus of Claims 1, 2, 3, 4, 5 and 6 , in which the arrays of input image display segments are formed of frames of TV images received from one or more TV channels, which are displayed on one or more TV display monitor units.

8 Apparatus of Claims 1 2, 3, 4, 5, 8 and 7, in which the input image display is generated by computer.

9 Apparatus of Claims 1, 2, 3, 4, 5, 6 and 7 , in which the means for displaying images onto one or more display surfaces is through the use of electroluminescent panel displays, or formed using coherent or incoherent tight sources.

Apparatus according to Claims 1, 2, 3, 4 and 5, in which means are provided for adjusting the lateral alignment of each particular page 35 lens with respect to the input image segment in front of it.

11. Apparatus according to Claims 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10, in which rows or columns of input image segments of each of a number of input images are individually mounted on the planar faces of mounting rods having a triangular, square, rectangular, or polygon-shaped cross-section, and means are provided to enable each of these rods to be synchronously rotated about their axis, which runs through their centroids, or centres, of their cross-section in their long dimension, through an angle chosen such that rows or columns of associated input image segments can be aligned with each other on a planar surface which is parallel to that on which the array of of lenses is located, with the result that a particular chosen output image will be reconstructed on an output image plane.

12 Apparatus of Claims 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 and 11, in which the lateral shapes of lenses are hexagonal, and the lenses are positioned to form regular or irregular close-packed arrays on planar surfaces, multifaceted surfaces, or curved surfaces which are touched by the centres of the lenses and the input image segments.

13 Apparatus of Claims 1, 2, 3 ,4, 5, 6, 7, 8, 9,10 and 11, in which the lateral shapes of lenses are triangular, and the lenses are positioned to form regular or irregular close-packed arrays on planar surfaces, multifaceted surfaces, or curved surfaces which are touched by the centres of the tenses and the input image segments.

14 Apparatus of Claims 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 and 11, in which the lateral shapes of lenses are square, and the lenses are positioned to form regular or irregular close-packed arrays on planar surfaces, multifaceted surfaces, or curved surfaces which are touched by the centres of the lenses and the input image segments.

15 Apparatus of Claims 1, 2, 3, 6, 7, 8, 9, 10, 11, 12, 13 and 14 , in which imagery of objects are recorded at a number of positions around, above or below the object, in order to obtain input imagery over a wide range of azimuthal and elevation angles, so that this imagery can subsequently be used to form image segments which are positioned with the centre of each input image segment and the lens behind it each touching their respective curved surface, or with one or a number of page 36 lenses and image segments on each of a number of planar facets of a multifaceted surface, which are convex towards viewers, such that as these viewers move around this convex viewing surface, they will observe magnified stereoscopic output imagery of these objects with different perspectives in different directions

16 Apparatus according to Claims 1, 2, 3, 4, 5, 6, 7, 8, 9,10,11, 12,13 and l4, in which the curved and the multifaceted surfaces on which arrays of lenses and input image segments are positioned are surfaces of icosahedra.

17 Apparatus according to Claims 1, 2_, 3, 4, 5, 6, 7, 8, 9,10,11, 12,13 and 14, in which the curved and the multifaceted surfaces on which the arrays of lenses and input image segments are positioned are surfaces of truncated icosehedra

18 Apparatus according to Claims 1, 2, 3, 4, 5, 6, 7, 8, 9,10,11, 12,13 and 14, in which the curved and the multifaceted surfaces on which the arrays of lenses and input image segments are positioned are surfaces of geodesic domes.

19 Apparatus according to Claims 1, 2, 3, 4, 5, 6, 7, 8, 9,10,11, 12,13 and 14, in which the curved and the multifaceted surfaces on which the arrays of lenses and input image segments are positioned are surfaces of ellipsoidal-shaped domes.

page 3 7