Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02447817 2003-10-15
METHOD AND APPARATUS FOR REFRACTIVE AND REFLECTIVE
POLYSTEREOSCOPIC IMAGING
Background of the Invention
Field of the Invention
The invention relates to polystereoscopic image capture and display
technology, and more particularly to methods and apparatus for capturing three
dimensional images and displaying the same to emulate motion.
Descriation of the Prior Art
Recent stereoscopic solutions have employed LCD panel displays with
lenticular lens sheets. Examples of this technology can be found in United
States
patent numbers 4,717,949; 4,829,365; and 6,157,424. However, physical size
constraints of LCD pixels and sub-pixels impose a number of constraints
including
the number of views (angular resolution), planar resolution per view, and
viewable
observer distance from screen
The invention disclosed in United States patent number 6,224,214 identifies a
plurality of projectors that cast images onto an active optical viewing
screen, which
employs shutters to control the projected angular views. In fact, any
embodiment
employing any kind of active light control mechanisms within the screen itself
is
limited in angular resolution as well as planar resolution due to the physical
limitations of how small the LCD or physical screen "shutters" can be
manufactured.
Other embodiments of this shutter technology such as United States patent
number 6,304,288 require head tracking to direct the appropriate stereoscopic
image
to the viewer's eyes, as determined by the viewer's location and orientation.
Multiple
viewers introduce a substantial complexity factor to this approach making it
unfeasible or extremely cost prohibitive for multiple viewer capabilities. The
same
restrictions apply to this approach as do to the above patent, i.e., the
6,224,214
patent. Active shutter size minimization constraints in turn restrict angular
resolution
and planar resolution.
1
CA 02447817 2003-10-15
Summary of the Invention
The purpose of the invention is to display, either simultaneously or
approximately simultaneously, a plurality of images onto a single surface from
a
plurality of discrete locations thereby providing one or more appropriately
positioned
viewers with a polystereoscopic image (one that has stereoscopic properties
from a
plurality of viewing locations). Thus, the invention comprises several related
components, in particular an image capturing component, a data translation
component, and an image display component.
Turning first to the image capturing component, a plurality (at least two) of
image capturing devices, such as still or motion video cameras, are positioned
so
that each device captures substantially the same target field but at unique
angles
with respect to such field. Preferably, the devices are located along a plane
that is
generally orthogonal to any primary objects located in the field and are
positioned in
an array. Each captured image In originates from a discrete image recording
device
(R), and comprises a plurality of discrete elements EX,y, which may be
considered
pixels, where R is unique (for convenience, R = an integer), n = the image
number, 0
< x < (maximum horizontal element count), and 0 < y < (maximum vertical
element
count). Each image (In) captured by each recording device (R) (referred to as
image
Rln) is stored at least temporarily for further use as will be described
below.
The data translation component of the invention comprises alternative means
for providing a plurality of image projection apparatus (P) with display data.
In part,
the selection of a suitable data translation component depends upon the means
available for displaying the captured images. It is well known that the human
eye
cannot easily discriminate between related images if the display rate is above
about
28 frames or images per second. Exploiting this deficiency, the objective of
the data
translation component is to present each image Rln within about 0.035 seconds
where each (R) is unique within the domain and (n) is constant. For example,
if RmaX
= 20, then a total of 560 images must be displayed in one second if n = 1 (20
images
2
CA 02447817 2003-10-15
x 28 images/second). The following second, another 560 images must be
displayed
when n = n + 1. Depending upon design considerations, several alternatives
might
be considered, e.g., a single projector many mirrors solution or a many
projectors
solution.
A first alternative is deemed a serial image display process and a second
alternative is deemed a serial element display process. In the first
alternative, each
image RI~ is sequentially written to a memory such as a display buffer, e.g.,
each
element EX,Y from one unique image RI" is sequentially written to a display
buffer
where (R) and n remain constant for the duration of the data writing. Once the
display buffer has been loaded, the unique image RI" is displayed, and the
buffer is
cleared or overwritten by data for the unique image (R+1)I~~~.~~.
In the second alternative, one element EX,y from each image Rln is
sequentially
written to a display buffer, where (R) = 1 ~ RmaX, n = 1, and x,y remain
constant.
Once the value (R) has reached its limit, the resultant image composition is
displayed
in a manner discussed in more detail below, and the buffer is cleared or
overwritten
by data for the next element EX+~,y+~ from each image Rln where (R) = 1 ->
Rmax and n
= n + 1. Selection of one projection arrangement over another may depend upon
external factors such as image compression algorithms, e.g., the change
between all
pixel elements "1" is less than the change of all pixels within a given image.
With respect to the foregoing, those persons skilled in the art will
appreciate
that the incrementing presented herein does not limit the invention, but
serves to
exemplify the operations of the invention. Thus, the incrementing can be via
multiples of integers other than 1 or the product of an applied optimization
algorithm.
Once an appropriate degree of image translation or conversion has taken
place, it is necessary to project the resolved stereoscopic images. Regardless
of the
mode of image translation, there are several means for projecting images to
the
selected screen. A first projection means involves a plurality of image
projectors that
3
CA 02447817 2003-10-15
create a projection matrix. Each projector is preferably positioned relative
to a target
surface in a manner similar to the location and orientation of the image
capturing
devices. A second projection means involves a plurality of mirrors that create
a matrix
and one or more projectors precisely aimed at the plurality of mirrors. In
this second
embodiment, each projector may project one or more images towards the mirrors,
and/or a rotatable mirror or refraction element may be positioned intermediate
one
projector and the mirror matrix to actively redirect a projected image.
Regardless of the projection means chosen, each projected image should be
precisely aligned relative to each other so as to cause superposition of each
projected image portion or pixel. It is to be noted that while each superposed
pixel
represents the same image object (if not interfered with by another object in
the light
path of that object to the image capturing device), the specific attributes of
that pixel
vary depending upon the angle of capture of that image pixel.
'I 5
According to the invention, there are two fundamental approaches to building
a parallax display that produces a stereoscopic image of remotely recorded
subject
matter. Refractive parallax display technology uses lenses to transmit an
image to a
"specific" viewing spot (can also be said in a specific viewing direction)
from a
corresponding location as emitted by a specific projector. Reflective parallax
display
technology uses directional reflection technologies to reflect light back to
its source of
origin. Both of these approaches employ related projection solutions to
resolve
commonly identified issues with existing parallax display technologies but
differ in the
"screen" technologies, insofar as one utilizes refraction while the other uses
reflection. Thus, once an image Rln has been projected, whether serial or
sequential,
the target surFace must be capable of either reflecting or refracting the
projected
image.
In a first embodiment, a refracting surface is used. The refracting surface
should be capable of receiving projected light from a plurality of incident
angles and
redirecting the incident light to emit it in a substantially parallel manner.
In this first
4
CA 02447817 2003-10-15
embodiment, a plurality of thin lenses formed in or on a planar surface
comprise the
refracting surface. The refracting surface serves two primary functions. The
first
function is to refocus the projected light onto a corresponding location in
the plane of
image convergence. Thus, the refracting surface functions as a large convex
lens
towards each projection passed through it. The second function relates the
conditioning of the individual pixels, which varies accarding to design
considerations
and the portion of the surface in which the beams of light are incident. Thus,
the
refracting surface can make one or more beams collimated (neither diverging
nor
converging); it can make one or more beams converge, either onto the plane of
convergence or other location as best deemed by empirical results; or it can
otherwise alter but maintain the diverging nature of the beams.
To achieve these primary functions, the refracting surface should have
qualities of a Fresnel lens employing individually unique conventional thin
lenses.
The refracting surface is preferably manufactured to include an integrated
lens
structure employing both lens types.
In the second embodiment, a directionally reflective (retroreflective) surface
is
used to reflect projected light generally back to the source or other target
location.
The directionally reflective surface comprises, in a preferred embodiment, a
plurality
of reflective elements formed in or embedded into a surface. The reflective
elements
are characterized as spherical or partially spherical badies that reflect
incident light
rays towards the source of such rays. A schematic representation of such
behavior
is shown in Fig. 9.
Because the screen elements are retroreflective, the issue of projector
opacity
must be addressed. The problem of projector opacity is that in a truly
retroreflective
environment, the incident and reflected beams are coincident. Thus, in order
to
observe the reflected beam, the observer must be in the incident beam path. If
the
observer is opaque, then the incident beam cannot transit to the reflector; if
the
projector is opaque, it will prevent observation of the reflected beam. To
overcome
this difficulty, in one embodiment a partial mirror or beam splitter is used
to redirect a
portion of the projected image towards the screen, and permit a portion of the
5
CA 02447817 2003-10-15
reflected image to pass through the mirror and be observed by the viewer(s).
The
partial mirror is positioned oblique to the viewing screen so as to receive
off angle
projection light and redirect the same to the viewing screen. Reflected light
is then
permitted to pass partially through the partial mirror back towards the
viewer(s).
Brief Description of the Drawings
Fig. 1 is a schematic diagram of a plurality of image capturing devices
exemplifying the stereoscopic recordation of an object by a pair of such
devices;
Fig. 2 illustrates a matrix conversion scheme illustrating the rearrangement
of
pixel projections from multiple projectors (top) to multi-angular refractive
or reflective
screen pixel emissions (bottom).
Fig. 3 is a schematic diagram illustrating sample beams of collimated light
emitted from two closely spaced projectors where the beams carry the same
information as was acquired by the cameras in Fig. 1;
Fig. 4A is an illustrative schematic view of a single convex lens, which is
approximated by the plurality of lenses in Fig.3, illustrating the desired
functionality of
a passive refractive display screen;
Fig. 4B is the same illustrative schematic view shown in Fig 4a, but the
projectors are arranged in an arc (if 3 dimensionally, a spherical cap) in
order to
remove the requirement from a macrolens that it provide a variable focal
length to
projectors located further from the center of the project array;
Fig. 5 is a schematic diagram illustrating the placement and projection range
of a plurality of linear projectors and convergence of beams of light
emanating from
each projector;
Fig. 6A shows a schematically isolated far lateral microlens receiving a
plurality of incident light beams originating from discrete angles of
incidence, and the
subsequent refraction of the incident light beams;
Fig. 6B shows a schematically isolated central microlens receiving a plurality
of incident light beams originating from discrete angles of incidence, and the
subsequent refraction of the incident light beams;
6
CA 02447817 2003-10-15
Fig. 6C is a schematic diagram illustrating a single light beam representing a
single pixel from a source entering into and exiting from the microlens shown
in Fig.
6a;
Fig. 6D is an illustrative schematic diagram of the discreet refractive
properties
of a single microlens wherein a shadow mask is used to limit incident beam
entrance
properties;
Fig. 6E is an illustrative schematic diagram of the discreet refractive
properties
of a single microlens using a compound lens structure to address unique
incident
beam angle refractions, thereby maintaining constant angular refraction
regardless of
incident angles;
Fig. 7A is a schematic diagram of an alternative projection arrangement
wherein a single projector and rotating mirror is used to sequentially target
a plurality
of discrete mirrors mounted to a surface;
Fig. 7B is a schematic diagram of an alternative projection arrangement
wherein a single projector and rotating mirror is used to sequentially target
a
continuous mirror;
Fig. 7C is a schematic diagram illustrating possible projector locations
necessary to achieve a desired refractive convergence when the refractive
screen
does not address suitable angular refraction;
Fig. 8 is a schematic diagram of a reflective display arrangement using a
partial mirror to permit a viewer to see a retroreflection;
Fig. 9 is a schematic diagram of a mircosphere used to achieve
retroreflection;
Fig. 10 is a perspective view of a reflective display screen;
Fig.11 is a detailed perspective view of a plurality of microspheres that
comprise the screen shown in Fig. 10; and
Fig. 12 is a side elevation taken substantially along the line 12-12 in Fig.
11
exemplifying the association of the microspheres and the screen backing.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
7
CA 02447817 2003-10-15
The following discussion is presented to enable a person skilled in the art to
make and use the invention. Various modifications to the preferred embodiment
will
be readily apparent to those skilled in the art, and the generic principles
herein may
be applied to other embodiments and applications without departing from the
spirit
and scope of the present invention as defined by the appended claims. Thus,
the
present invention is not intended to be limited to the embodiment shown, but
is to be
accorded the widest scope consistent with the principles and features
disclosed
herein.
Turning then to the several Figures wherein like numerals indicate like
components, and more particularly to Figs. 1-2 and 3-7, the image acquisition
and
passive refractive display features are illustrated, respectively. It should
be noted
that for illustration purposes, all drawings and descriptions herein (unless
otherwise
noted) are directed to a single line or horizontal sweep (i.e., "x" coordinate
axis). It is
to be understood that for "y" coordinates to be captured andlor displayed,
repetition
of the illustrated embodiments along the "y" axis must be carried out, as is
contemplated in a preferred embodiment.
Image Acquisition
In order for the invention to operate, appropriate image information must be
obtained. Turning then to Fig. 1, a one-dimensional array of image recording
devices
in the form of digital cameras 22, 24 mounted to plane 20 is shown. It is to
be
understood that while the functioning of only two cameras are described, a
plurality of
cameras is used as is illustrated in this Figure. Camera 22 captures an image
(I) of
an object bounded by the leftlright image plane 30 coordinates 26, 28 to
define its
field of view and is preferably aimed directly at (and presumably focused on)
the
object from its unique vantage point. Thus, the captured image (I) from camera
22
can be identified as 221. In the same manner, camera 24 captures an image (I)
of the
object bounded by the leftlright coordinates 26, 28 to define its field of
view and is
also preferably aimed directly at (and presumably focused on) the object from
its
unique vantage point. Thus, the captured image (I) from camera 24 can be
identified
as 241. The distance between cameras 22 and 24 is represented by the distance
"x".
The value "x" represents the horizontal stereoscopic resolution of the
composite
8
CA 02447817 2003-10-15
image (I) being captured. This camera positioning scheme is repeated for all
cameras mounted to plane 20.
An alternative to this "many camera" scheme is to use a discrete mirror or
contiguous mirror arrangement to replace the plurality of required cameras.
Instead,
a passive mirror array or mechanism covering the necessary vantage points
could
reflect the object from each specific vantage point onto a rotating prism or
mirror
mechanism, which then redirects the reflection into a high-scan rate image
capturing
device. This arrangement is complementary to the projection arrangement shown
in
Figs. 7A and 7B.
Depending on the artistic intentions of the image acquisition setup, captured
scenes may require that all information in the scene be in focus. Such a focus
requirement presents a problem, as each camera lens typically has one focal
surface
which is usually set to pass through the object being recorded. The image
acquisition requires that all foreground and background objects be in focus as
captured by each camera. One way to achieve this task is to use wide-field
lenses
on the camera array. Wide-field lenses tend to keep all imaged objects in
focus, but
also disproportionately distort the size of objects that are closer to the
camera.
Another way is to minimize the size of the camera aperture a sufficiently
large depth
of field is achieved. Again, there are out-of focus distortions that result
from this
solution which occur on the fringes of the image. Both means require
additional
computer processing to electronically correct distortions by reparameterizing
the light
field as captured by the multiple camera vantage points to provide multiple
focal
surfaces and ensure that all objects are in focus, as discussed in (Isaksen,
Aaron,
Leonard McMillan, and Steven J. Gortler. Dynamically Reparameterized Light
Fields.
SIGGRAPH 2000), which is incorporated herein by reference. Yet another
solution is
to ensure that a minimum operating distance between the subject matter and
surrounding scenery is maintained in order to empirically avoid distortion
issues
introduced by wide-field lenses or high depth of field camera implementations.
Image Data Manipulation
9
CA 02447817 2003-10-15
In a preferred embodiment, each camera captures a plurality of single frames
(representing a plurality of single images (I") where (n) represents a unique
frame or
image number), which are stored on a time indexed video tape, magnetic storage
media, optical disc, or similar medium. If not stored on a digital format the
data
comprising the plurality of single frames is converted into digital format or
stored on
standard high speed photographic film.
For each frame of data, each pixel therein is assigned a unique address.
Thus, an image (In) originating from camera 22 is labeled 221, and has a
plurality of
pixel elements EX,y (the x,y naming convention is useful for a two dimensional
array,
but serves to exemplify the naming conventions for the invention; for the
purposes of
illustration herein, only the "x" coordinate designation is used); this
results in a unique
address for each pixel data of 221nEx,Y. Similarly, pixel data for an image
originating
from camera 24 would be represented as 24fnEX,y. For purposes of this
disclosure,
each image (I) comprises only three pixel elements, namely Ea, Eb, and E~. A
database of all image data is then created for subsequent use as will now be
described.
Returning to Fig. 1, it becomes apparent that camera 22 captured three
discrete pixels of data: 221~Ea, 221~Eb, and 221~E~. Similarly, camera 24
captured
three discrete pixels of data: 241~Ea, 24f~Eb, and 241~E~. As is illustrated
in Fig. 3, this
data is converted by projectors 42 and 44 into corresponding beams of light:
beams
42a corresponding to 221~Ea, 42b corresponding to 221~Eb and 421c
corresponding to
22~E~ from projector 42; and beams 44a corresponding to 241~Ea, 44b
corresponding
to 241~Eb and 44c corresponding to 241~E~ from projector 44. It should be
noted that
projectors 42 and 44 are positioned identically with respect to refractive
screen 50 as
were cameras 22 and 24 with respect to image plane 30 so as to maintain the
fidelity
of image reproduction.
CA 02447817 2003-10-15
As will be discussed in more detail below, the labeling of each pixel permits
various display options, including generating composite images using pixels
derived
from various image capturing devices.
Implicit to the invention is the reorganization of the data from each
projector
(or projector vantage point) to the screen (reflective and refractive). Fig 2
shows 8
projectors 10-18 each projecting an image comprising of 3 pixels, a, b, c.
Through
the crisscrossing of the pixel light beams, the data is rearrange so as to be
re-emitted
by 3 pixels on the reflective/refractive screen. Each screen pixel emits light
in 18
different directions so that the pixel is perceived as having a quality, which
depends
on the viewer's vantage point.
Substantial cost reduction is achieved in that the signals as emitted by
common, off-the-shelf projectors) are standard image projections, which are
converted by the invention arrangement for re-emission by the refractive or
reflective
screen. This matrix is special in that it describes a process native to the
technology
that would otherwise involve substantial processing requirements or active
screen
components such LCD shutters. As well, in utilizing a passive optical screen
in both
the refractive and reflective modes, this matrix is a logical-pixel-
rearrangement
example of how a multitude of angular views for each passive optical "pixel"
is
effectively achieved.
Refractive Display
In order to achieve the beam redirection shown in Fig. 3, refractive screen 50
comprises a plurality of microlenses 52, each laterally unique one from the
other.
The purpose of refractive screen 50 is to approximate a single macrolens with
a
plurality of unique optical properties. Each unique property is only evident
to a
specific beam path such that different light beams from different projectors
are
affected by the macrolens as if there were a different macrolens with
different
properties for each projector. For example, projections coming from projector
42
along path 42a are refracted by microlens 54, which provides a longer focal
length
11
CA 02447817 2003-10-15
than projections coming from projector 46. An enlarged functional illustration
of a
pair of macrolenses 50' and 50" is shown in Fig. 4. As shown therein, the same
beams of light 42a, 42b, 42c, 46a, 46b, and 46c are redirected to focal plane
30' as
was the case in Fig. 3. Consequently, refractive screen 50 in Fig. 3 has a
functionality equivalent to macrolenses 50' and 50" for these beam paths. A
potential
simplification of the varying focal length requirements placed upon macrolens
50' is
to arrange the projectors in a spherical cap instead of a plane as in Fig. 4a.
In this
arrangement, the focal lengths of macrolens 50' and 50" are identical.
Projectors
may be set in other arrangements to further minimize other optical
requirements of
the refractive screen.
Those persons skilled in the art will appreciate that the focal length between
beam 42a and 42c is different. Thus, unless corrective optics are employed,
the
entire projected image on the back of the refractive screen 50 will not be in
focus with
respect to distally lateral projectors. This correction may ar may not be
required,
depending on the specific refractive screen implementation. The correction can
take
place anywhere along the incident beam path, such as in the projector optics,
beam
splitter, or rotating reflector (discussed below). Alternatively, the geometry
of surface
40 and/or refractive screen 50 can be altered to maintain a constant focal
length
between the plurality of projectors and the screen.
Those persons skilled in the art will also appreciate that a laterally distal
projector such as projector 42 will cause a skewed image to appear on
refractive
screen 50 as compared to a central projector such as projector 46. To overcome
this
optical deficiency, the displayed image can either be digitally corrected via
keystone
adjustment, or can be optically corrected by mounting a unique lens assembly
to the
projector. In either solution, the objective is to acquire a true image (i.e.,
one that has
the same optical properties as a centrally projected image).
Returning then to refractive screen 50 in Fig. 5, each microlens 52 is
laterally
unique, one from another. In this manner, the intended directional orientation
of
12
CA 02447817 2003-10-15
refracted light is maintained regardless of its source. Figures 6A-B
schematically
illustrate the beam redirection properties of a distally lateral microlens
(for illustrative
purposes, a thin lens is shown while a preferred embodiment would use a
Fresnel
fens having the same or highly similar beam redirection properties) in 6A, and
a
central microlens in 6B. As illustrated in Fig. 6A, each incident beam is
refracted by a
constant amount. As illustrated in Fig. 6B, each incident beam is nominally
refracted.
Moreover and as functionally illustrated in Fig. 6C, beam divergence, which is
an artifact of the projected image (as the distance between the image plane
and the
projector increases, so does the image size), is addressed insofar as the
refracted
beam is either conditioned to converge or maintain a parallel profile. The
beam may
be set to converge, maintain a parallel profile or even diverge to an extent
in order to
help blend one vantage point into another as the observer moves wigthin the
viewing
space. If all beams emitted by all microlenses as projected by all projectors
converge
onto a singe point, the refractive screen will appear black for viewing
locations not
specifically targeted by a camera vantage points. A parallel profile or
divergent
profile on the beam from each microlens mitigates this by "diffusing" the
image so
that it is still visible from viewing locations in the immediate vicinity of
the specifically
targeted vantage point. The functionally flat input surface on each microlens
gives
the refracting screen surface the properties of a convex macrolens (and is
illustrated
as such} by redirecting the beam in a new direction. The exit surface of the
microlens controls the convergence/divergence properties of the light beam.
The
exiting beam in Fig. 6C is shown as perfectly collimated. Furthermore, the
incident
surface does not have to be flat, but may have a specifically calculated
functional
curvature as required by the location andlor other properties of the projector
array.
See Fig. 6E. The relation between the angle of the microlens entry surface and
the
screen wilt vary, depending of the location of the microlens on the screen,
e.g., at the
center of the screen, the angle (if a flat entry surface is present) would be
0 degrees.
While for simplicity there is an implied 1:1 correlation between projected
pixel beams
and corresponding microlenses, the correlation does not have to be the same,
given
that the lenses can be sufficiently miniaturized such that one pixel beam hits
2, 3 or
13
CA 02447817 2003-10-15
more lenses simultaneously. By eliminating the 1:1 correlation, certain
manufacturing and alignment costs can be substantially reduced.
Figure 6D is a functional schematic diagram of one of the possible microlens
structures that would address the "angle constant requirement" that each
microlens,
as described in Fig. 6A and Fig. 6C, should have. Specially, Fig. 6D
illustrates a
grouping of hypothetical sub-microlenses 52a, 52b, and 52c, which together
make up
one of the primary colors (see below) of one pixel on the refracting screen.
Each
sub-microlens 52a-c is exposed to light from only a single direction, as
determined by
slit 62 in shadow mask 60. Shadow mask 60 preferably covers the incident side
of
the refractive screen. In this manner, each projected incident beam 48a-c is
conditioned by a slit 62 and uniquely (or approximately uniquely) refracted by
a
microlens 52 to produce refracted beams 49a-c wherein the total angle of
refraction
(indicated by arrows) is the same or substantially the same.
The continued extension of the discrete sub-pixel micro-lenses shown
illustratively in Fig 6D is a single microlens 52' with a specifically
calculated entry
surface as Fig 6E demonstrates, to function as the individual sub-pixel
microlens in
Fig 6D. The two important factors to Fig 6E are that the angle of incidence of
incoming light and the tangent at the point of entry into the microlens is the
same for
all incident beams, 48a, 48b, and 48c. In like manner, all refracted beams
exiting
microlens 52' have a perpendicular exit path relative to microlens 52'.
While not illustrated herein, any differences in angle of refraction between
the
three primary colors, i.e., red, green and blue, are rectified by having three
hypothetical microlenses as defined in Fig 6E for each pixel coordinate
targeted by all
projectors, and three color-filtered slits in the shadow mask there at. All
projectors in
the array will, for example, send pixel (1,1 ) to the same coordinate on the
refractive
screen. At that location, there are 3 color-filtered slits in the shadow mask
allowing
each of the RGB colors to pass through on to the respective hypothetical
microlens.
Each hypothetical microlens will have a specific entry surface curvature to
property
14
CA 02447817 2003-10-15
refract light coming from the specific directions as determined by the slit.
Again, the
properties of each hypothetical microlens are preferably incorporated into a
fewer
number of functional composite lenses such as the type illustrated in Fig. 6E.
Heretofore, this embodiment of the invention has been described in terms of
multiple projectors, e.g., one projector for each image capturing device.
However, it
is contemplated to use a single projector for a plurality of image capturing
devices.
Turning attention to Figs. 7A and 7B, it can be seen that the output of a
single high-
scan rate projector 70 can be directed to rotating reflectorlrefractor 80,
which in turn
distributes a projected image to reflective member 90, which may be mirror
array 92
having a plurality of discrete reflective surfaces 94 as shown in Fig. 7, or
may be
continuous reflective member 96 as shown in Fig. 7B. By rapidly sequencing a
plurality of discrete projector images during operation of projector 70 and
synchronizing rotation of reflector/refractor 80 to deliver each image to a
unique
position on reflective member 90, a fewer number of projectors, or even a
single
projector can be used.
It is well known that in order to emulate motion, it is necessary to have a
frame
rate in excess of about 25 FPS. Consequently, if a projector has a vertical
refresh
rate of 150 Hz, it is capable of displaying six (6) unique image (frame)
sequences per
second and still maintain motion emulation. In the illustrated embodiment, if
projector
70 has a vertical refresh rate of 150 Hz, then reflectorlrefractor 80 may have
6 facets
and will therefore redirect 6 images to screen 50 during a single rotation.
Naturally,
the relative rotation rate and optical characteristics of reflector/refractor
80 may vary
depending upon overall design considerations, and a rate of 50 rpms has been
chosen for illustration purposes only. A benefit of this approach is that for
similarly
grouped projectors (ones that have substantially the same focal plane and
keystone
correction factor) this expedient can be applied without notably degrading the
image
quality. Moreover, consistency issues relating to minor dififerences in
projectors can
be eliminated by using a single projector for providing the imaging to
multiple,
discrete projector locations.
CA 02447817 2003-10-15
Numerous variations on this approach can be employed. For example, it is
contemplated to modify the spatial geometry of reflective member 90, utilize
refraction principles in rotating reflectorlrefractor 80 so as to address
optical issues,
and distribute projection images in the "y" axis.
Of the current digital projector technologies, the Digital Light Processing
("DLP") technology developed and marketed by Texas Instruments offers vertical
refresh capabilities to potentially enable a single projector to provide all
necessary
frames from all required vantage points via rotating reflector/refractor 80.
However,
the DLP technology Digital Micromirror Device introduces time as a variable in
composing the projected image. A minimum time period is required to carry out
the
binary pulse-width modulation for each projected frame in order to construct
the
varying shades of the frame's RGB components. Three mechanisms may be
employed individually or in tandem to mitigate time synchronization
incompatibilities
between the rotating reflectorlrefractor 80 and the Digital Micromirror
Device. The
rotating reflector/refractor can be driven by a stepper motor, able to stop in
specific
positions, so that the rotation is comprised of discrete reflectorlrefractor
movements
instead of a continuous rotational sweep. If a continuous rotational sweep is
desired,
a plurality of rotating reflectors/refractors can be set to work in unison to
keep the
image specifically aligned on a specific reflector of reflective member 90
without
halting the rotational momentum of the reflectors/refractors. The third method
is to
increase the vertical refresh rate of the DMD chip in order to decrease the
time-
period required by the DLP projector to construct all necessary shades of each
frame's RGB components.
Heretofore, features of the invention addressed the issue of incident beam
angles and constant angle refraction so that the refracted beams would
properly
converge on the focal plane. Special optical properties were employed to
address
extreme lateral projections, variable focal lengths and the like. However, an
alternative scheme is proposed that would significantly reduce the requirement
for
optical solutions to these issues. Figure 7C illustrates a situation wherein
comparatively homogenous microlenses 52' are used throughout refractive screen
16
CA 02447817 2003-10-15
52". Since these microlenses are not of the "angle constant" type (a rather
unique
microlens for each screen location), incident beams 48a-c do not emanate from
a
single location on projector array plane 40 in order to converge on focal plan
30'.
Because of the nature of image processing as described above and illustrated
in Fig.
2, it is possible to produce a composite image for each projection location.
Thus, an
image projected from projector 42 in Fig. 3 may comprise pixel data from a
plurality of
cameras. Returning to Fig. 7C, incident beam 48a would be emanate from a
projector located at point "a" on plane 40, which includes corresponding image
data
acquired from a similarly positioned camera; incident beams 48b and 48c would
emanate from a projector located at point "b-c", which includes corresponding
image
data acquired from a similarly positioned camera. However, even though the
projected beams emanate from discrete locations, they converge at a common
location after refraction by screen 52". Thus, deficiencies or expedients
regarding
screen 50 or 50" may be addressed by modifying the pixel projections for any
given
image (I) being projected. Since the process in Fig. 2 illustrates that each
pixel for
any given image may be projected in any selected manner, the unique projection
locations for each image (I) can be determined and implemented via a suitable
algorithm.
Reflective Display
Previously in this disclosure, the means for viewing the acquired image data
was by way of projection onto a refractive screen. However as illustrated in
Figs. 8-
12, the invention also includes a passive reflective display implementation.
In this
implementation, retroreflection is used to provide a viewer with stereographic
imaging. Through experimentation, it has been verified that microspheres
provide
the desired means for achieving retroreflection. Tests have successfully been
conducted using the following products: 3M~ Scotchlite~ black retroreflective
film;
Swarco Megalux-Beads~.
A highly desirable property of microspheres is their ability to directionally
reflect emitted light back to the light origin. This retroreflectivity is a
desirable
component to the implementation of this feature of the invention. Fig. 9
illustrates the
17
CA 02447817 2003-10-15
retroreflectivity of a single microsphere 100. This arrangement advantageously
addresses the issue of beam divergence that would otherwise occur with a non-
retroreffective surface, such as a planar polished surface, which maintains
the beam
divergence property. As shown in this Figure, exiting light is either
collimated or
converging; if diverging beams of fight enter the sphere, they are reflected
so as to
generally mimic the emission path of the light beam source. Thus, if a
projector is
located at an optical distance "z" from the screen, a viewer similarly located
would
perceive a sharp and accurate image as projected on the screen. In addition,
lateral
resolution is maintained as reflection bleed (one pixel to the next) is
minimized.
To construct passive reflective screen 150, a plurality of microspheres 100
are
mounted to a suitable planar surface as is illustrated in Figs. 10-12. To
minimize
reflection of incident or ambient light, adhesive 152, which serves to bind
microspheres 100 to backing member 154, should have very low reflective
properties; alternatively, a separate coating can be applied over screen 150
where
after the coating is removed from spheres 100 but not adhesive 152. If using a
coating solution, the solution should not significantly bind to the spheres so
that the
coating can be removed from the spheres but not from the adhesive.
Also when utilizing a screen as described herein and illustrated in Figs. 10-
12,
true retroreflection will result in the reflected fight being redirected to
the projectors.
A viewer must be positioned so as to intercept the reflected light in order to
perceive
the projected image, however, such a viewer would then also interfere with the
projected light. Consequently, provisions must be made to permit simultaneous
projection and observation. Figure 8 illustrates a means for accomplishing
this
objective. fn particular, a single projector 146 is functionally shown
directing light
beams 148 towards angled half mirror or beam splitter 160. A portion of the
projected light beams 148 reflects from beam splitter 160 as light beams 248
and
impinges on screen 150. Screen 150 in turn retroreflects light beams back to
beam
splitter 160. A portion of this retroreflected light 249 passes through beam
splitter
160, where after it enters the observer space. While the light observed by a
viewer
18
CA 02447817 2003-10-15
has an illumination of 25% of the projected source light, this solution
satisfactorily
addresses the issue of projection beam obstruction.
As was the case with the refractive screen discussed above, it is important to
align the multiple projected images so that each image is superposed upon each
other, i.e., each image pixel from each projector is substantially projected
to the same
location of the screen. In this manner, any given area on the screen will
represent
the same object, but from a plurality of viewing angles.
19