 |
Claims  |
|
|
We claim:
1. A method of digitizing an item, said method comprising the steps of:
(a) operating a first video camera to effect producing camera coordinates
of first and second alignment points in the field of view of said first
video camera and to effect producing a first plurality of camera
coordinates of a third alignment point, said third alignment point also
being in the field of view of a second video camera;
(b) operating said second video camera to effect producing a second group
of camera coordinates of said third alignment point, and to effect
producing camera coordinates of a fourth alignment point in the field of
view of said second video camera;
(c) computing skew corrected coordinates of said first alignment point from
said camera coordinates of said first and second alignment points, said
skew corrected coordinates of said first alignment point being referenced
to a boundary of a skew corrected camera plane of said first video camera;
(d) computing a first plurality of skew corrected coordinates of said third
alignment point from said camera coordinates of said third alignment point
and said skew corrected coordinates of said first alignment point, said
skew corrected coordinates of said third alignment point being referenced
to said first alignment point;
(e) computing skew corrected, offset corrected coordinates of said third
alignment point from both said skew corrected coordinates of said third
alignment point and an offset number representing the distance of a
boundary of a document plane to the boundary of said skew corrected camera
plane of said first video camera, said skew corrected, offset corrected
coordinates of said third alignment point being referenced to the boundary
of said document plane;
(f) computing a second plurality of skew corrected coordinates of said
third alignment point from said second group of camera coordinates of said
third alignment point and said camera coordinates of said fourth alignment
point, said second plurality of skew corrected coordinates of said third
alignment point being referenced to a boundary of a skew coorected plane
of said second video camera;
(g) operating said second video camera to effect producing of camera
coordinates of a general point in the field of view of said second video
camera, said general point being a point of the item to be digitized;
(h) computing skew corrected coordinates of said general point from said
camera coordinates of said general point and said second plurality of skew
corrected coordinates of said third alignment point, said skew corrected
coordinates of said general point being referenced to said third alignment
point;
(i) computing skew corrected, offset corrected coordinates of said general
point from said skew corrected coordinates of said general point and said
skew corrected, offset corrected coordinates of said third alignment
point, said skew corrected, offset corrected coordinates of said general
point being referenced to said boundary of said document plane;
(j) outputting said skew corrected, offset corrected coordinates of said
general point as a digitized value representing the location of said
general point on said document.
2. The method of claim 1 wherein said first, second, third and fourth
alignment points are disposed on a support surface that supports the item
to be digitized and wherein said method includes placing the item to be
digitized against the support surface between the support surface and said
first and second video cameras after step (b) and before step (g).
3. The method of claim 2 wherein the item to be digitized is a document,
the document covering said first, second, third and fourth alignment marks
when the document is placed on the support surface.
4. The method of claim 3 including the method of storing said camera
coordinates of said first, second, third and fourth alignment points.
5. The method of claim 4 wherein steps (c), (d), (e), (f), (h) and (i) are
effectuated by operating a microprocessor system.
6. The method of claim 5 wherein steps (a), (b) and (g) are also
effectuated by operating the microprocessor system.
7. The method of claim 1 wherein step (c) includes computing said skew
corrected coordinates of said first alignment point in accordance with the
formulas
x'.sub.A =kx.sub.A and y'.sub.A =ky.sub.A
wherein x.sub.A and y.sub.A are the camera coordinates of said first video
camera for said first alignment point and k is the ratio of the distance
between said first and second alignment points and the difference between
a y camera coordinate of said second alignment point and a y camera
coordinate of said first alignment point.
8. The method of claim 7 wherein step (d) includes computing said first
plurality of skew corrected coordinates of said third alignment point in
accordance with the formulas
x.sub.P '"=(x.sub.P -x.sub.A) cos .psi.+(y.sub.P -y.sub.A) sine .psi.
and
y.sub.P '"=-(x.sub.P -x.sub.A) sine .psi.+(y.sub.P -y.sub.A) cos .psi.
where x.sub.P and y.sub.P are the camera coordinates of said first video
camera for said third alignment point, and .psi. is the angle by which
said first video camera is skewed.
9. The method of claim 8 wherein step (e) includes computing said skew
corrected, offset corrected coordinates of said third alignment point in
accordance with the formulas
x.sub.P ""=x.sub.P "'+x.sub.PD '
and
y.sub.P ""=y.sub.P "'+y.sub.PD '
where
x'.sub.PD is an offset number representing the distance between an x
boundary of said document plane and said first alignment point and
y'.sub.PD is an offset number representing the distance between a y
boundary of said document plane and said first alignment point.
10. The method of claim 9 wherein step (f) includes computing said second
plurality of skew corrected coordinates of said third alignment point in
accordance with the formulas
x.sub.P '=k.sub.2 x.sub.P' and y.sub.P '=k.sub.2 y.sub.P'
where
x.sub.p' and y.sub.P' are camera coordinates of said second video camera
for said third alignment point and
k.sub.2 is the ratio between the distance between said third alignment
point and a fourth alignment point and the difference between a y camera
coordinate of said second video camera for said fourth alignment point and
a y camera coordinate of said second video camera for said third alignment
point.
11. The method of claim 10 wherein step (h) includes computing said skew
corrected coordinates of said general point in accordance with the
formulas
x.sub.G '"=(x.sub.G -x.sub.P') cos .psi..sub.2 +(y.sub.G -y.sub.P') sine
.psi..sub.2 and
y.sub.G '"=-(x.sub.G -x.sub.P) sine .psi..sub.2 +(y.sub.G -y.sub.P') cos
.psi..sub.2
where x.sub.G and y.sub.G are camera coordinates of said second video
camera for said general point, and .psi..sub.2 is the skew angle of said
second video camera.
12. The method of claim 11 wherein step (i) includes computing said skew
corrected, offset corrected coordinates of said general point in
accordance with the formulas
x.sub.G ""=x.sub.G "'+x.sub.GD '
y.sub.G ""=y.sub.G "'+y.sub.GD '
where x.sub.GD ' is an offset number representing the distance between said
x boundary of said document plane and said third alignment point and
y.sub.GD ' is an offset number representing the distance between said y
boundary of said document plane and said third alignment point.
13. The method of claim 12 including repeating steps (g) through (i) for
all other general points on said document in the field of view of said
second video camera.
14. The method of claim 1 wherein said first, second, third and fourth
alignment points are on the item to be digitized.
15. The method of claim 14 including the steps of producing codes to
represent different degrees of darkness of points scanned by said first
and second video cameras, said first, second, third and fourth alignment
points having a degree of darkness that is recognized for the purposes of
steps (a) (b), (c), (d), (e) and (g) but is not recognized for the
purposes of steps (h) and (i) to cause digitizing of all general points on
said document and avoid digitizing alignment marks.
16. The method of claim 1 wherein said first alignment point is located in
the upper left corner portion of the field of view of said first video
camera and said third alignment point is located in the upper left corner
portion of said second video camera.
17. A system for digitizing an item, said digitizing comprising in
combination:
(a) first and second video cameras;
(b) first means for operating a first video camera to effect producing
camera coordinates of first and second alignment points in the field of
view of said first video camera and to effect producing a first plurality
of camera coordinates of a third alignment point, said third alignment
point also being in the field of view of a second video camera;
(c) second means for operating said second video camera to effect producing
a second group of camera coordinates of said third alignment point, and to
effect producing camera coordinates of a fourth alignment point in the
field of view of said second video camera;
(d) third means for computing skew corrected coordinates of said first
alignment point from said camera coordinates of said first and second
alignment points, said skew corrected coordinates of said first alignment
point being referenced to a boundary of a skew corrected camera plane of
said first video camera;
(e) fourth means for computing a first plurality of skew corrected
coodinates of said third alignment point from said camera coordinates of
said third alignment point and said skew corrected coordinates of said
first alignment point, said skew corrected coordinates of said third
alignment point being referenced to said first alignment point;
(f) fifth means for computing skew corrected, offset corrected coordinates
of said third alignment point from both said skew corrected coordinates of
said third alignment point and an offset number representing the distance
of a boundary of a document plane to the boundary of said skew corrected
camera plane of said first video camera, said skew corrected, offset
corrected coordinates of said third alignment point being referenced to
the boundary of said document plane;
(g) sixth means for computing a second plurality of skew corrected
coordinates of said third alignment point from said second group of camera
coordinates of said third alignment point and said camera coordinates of
said fourth alignment point, said second plurality of skew corrected
coordinates of said third alignment point being referenced to a boundary
of a skew corrected plane of said second video camera;
(h) seventh means for operating said second video camera to effect
producing of camera coordinates of a general point in the field of view of
said second video camera;
(i) eighth means for computing skew coordinates of said general point from
said camera coordinates of said general point and said second plurality of
skew corrected coordinates of said third alignment point, said skew
corrected coordinates of said general point being referenced to said third
alignment point;
(j) ninth means for computing skew corrected, offset corrected coordinates
of said general point from said skew corrected coordinates of said general
point and said skew corrected, offset corrected coordinates of said third
alignment point, said skew corrected, offset corrected coordinates of said
general point being referenced to said boundary of said document plane;
and
(k) means for outputting said skew corrected, offset corrected coordinates
of said general point as a digitized value representing the location of
said general point on said item to be digitized.
18. The digitizing system of claim 17 wherein said first, second, third and
fourth alignment points are disposed on a support surface that supports
the item to be digitized.
19. The digitizing system of claim 18 wherein the item to be digitized is a
document, the document covering said first, second, third and fourth
alignment marks when said document is placed on the support surface.
20. The digitizing system of claim 19 including means for storing said
camera coordinates of said first, second, third and fourth alignment
points.
21. The digitizing system of claim 20 wherein said third, fourth, fifth,
sixth, eighth and ninth means are implemented by means of a microprocessor
system.
22. The digitizing system of claim 21 wherein said first, second, and
seventh means are implemented by means of said microprocessor system.
23. The digitizing system of claim 17 wherein said third means computes
said skew corrected coordinates of said first alignment point in
accordance with the formulas
x'.sub.A +kx.sub.A and y'.sub.A =ky.sub.A
wherein x.sub.A and y.sub.A are the camera coordinates of said first
alignment point and k is the ratio of the distance between said first and
second alignment points and the difference between a y camera coordinate
of said second alignment point and a y camera coordinate of said first
alignment point.
24. The digitizing system of claim 23 wherein said fourth means computes
said first plurality of skew corrected coordinates of said third alignment
point in accordance with the formulas
x.sub.P '"=(x.sub.P -x.sub.A) cos .psi.+(y.sub.P -y.sub.A) sine .psi.
y.sub.P '"=-(x.sub.P -x.sub.A) sine .psi.+(y.sub.P -y.sub.A) cos .psi.
where x.sub.P and y.sub.P are the camera coordinates of said third
alignment point, and .psi. is the angle by which said first video camera
is skewed.
25. The digitizing system of claim 24 wherein said fifth means computes
said skew corrected, offset corrected coordinates of said third alignment
point in accordance with the formulas
x.sub.P ""=x.sub.P "'+x.sub.PD '
and
y.sub.P ""=y.sub.P "'+y.sub.PD '
where x'.sub.PD is an offset number representing the distance between a x
boundary of said document plane and said first alignment point and
y'.sub.PD is an offset number representing the distance between a y
boundary of said document plane and said first alignment point.
26. The digitizing method of claim 25 wherein said sixth means computes
said second plurality of skew corrected coordinates of said third
alignment point in accordance with the formulas
x.sub.P '=k.sub.2 x.sub.P' and y.sub.P '=k.sub.2 y.sub.P'
where
x.sub.P', and y.sub.P' are camera coordinates of said video camera for said
third alignment point and
k.sub.2 is the ratio between the distance between said third alignment
point and a fourth alignment point and the difference between a y camera
coordinate of said second video camera for said fourth alignment point and
a y camera coordinate of said second video camera for said third alignment
point.
27. The digitizing system of claim 26 wherein said eighth means computes
said skew corrected coordinates of said general point in accordance with
the formulas
x.sub.G '"=(x.sub.G -x.sub.P') cos .psi..sub.2 +(y.sub.G -x.sub.P') sine
.psi..sub.2 and
y.sub.G '"=-(x.sub.G -x.sub.P') sine .psi..sub.2 +(y.sub.G -y.sub.P') cos
.psi..sub.2
where x.sub.G and y.sub.G are camera coordinates of said second video
camera for said general point, and .psi..sub.2 is the skew angle of said
second video camera.
28. The digitizing system of claim 27 wherein said ninth means computes
said skew corrected, offset corrected coordinates of said general point in
accordance with the formulas
x.sub.G ""=x.sub.G "'+x.sub.GD '
y.sub.G ""=y.sub.G "'+y.sub.GD '
where x.sub.GD ' is an offset number representing the distance between said
x boundary of said document plane and an x plane of said skew corrected
plane of said second video camera and y.sub.GD ' is an offset number
representing the distance between said y boundary of said document plane
and a y plane of said skew corrected plane of said second video camera.
29. The digitizing system of claim 17 wherein said first, second, third and
fourth alignment points are on the item to be digitized.
30. The digitizing system of claim 29 including means for producing codes
to represent different degrees of darkness of points scanned by said first
and second video cameras, said first, second, third and fourth alignment
points having a degree of darkness that is recognized by said first,
second, third, fourth, sixth and seven means but is not recognized by said
eighth and ninth means to cause said digitizing system to digitize all
general points on said documents but to not digitize alignment points.
31. The digitizing system of claim 17 wherein said first alignment point is
located in the upper left corner portion of the field of view of said
first video camera and said third alignment point is located in the upper
left corner portion of said second video camera.
32. The digitizing system of claim 17 including means for causing a
reference surface to have a predetermined level of brightness, means for
causing one of said video cameras to scan a first predetermined point of
said reference surface, analog-to-digital conversion means for producing a
reference code representing the brightness of said first predetermined
point, means for storing said reference code, means for causing the other
of said video cameras to scan a second predetermined point of said
reference surface, means for causing said analog-to-digital conversion
means to produce a brightness code representing the brightness of said
second predetermined point, means for comparing said brightness code and
said reference code, to produce adjustment information representing an
amount of adjustment required to adjust said analog-to-digital conversion
means to cause said analog-to-digital conversion means to produce a new
value of said brightness code that is more nearly equal to said reference
code, means for storing said adjustment information, and means for
adjusting said analog-to-digital conversion means in response to said
adjustment information. |
|
 |
|
 |
Claims |
|
|
|
Description |
|
|
BACKGROUND OF THE INVENTION
The invention relates to low cost, automatic digitizers and methods for
rapidly digitizing documents, and particularly to devices and methods
utilizing multiple video cameras to achieve rapid automatic digitizing of
large documents that extend beyond the field of view of a single video
camera.
Various types of digitizers, including manual digitizers, are well known.
Manual digitizers are widely used for low cost digitizing applications.
Manual digitizers are those that require manual control of the positioning
of a cursor or electronic pen or the like on a digitizing surface adjacent
to a desired point of a document. Manual digitizers convert the
coordinates of the desired point to digital numbers, either in response to
a "digitize command" or continuously, as an electronic pen or cursor is
positioned at a point of, or is removed along a line of, the document.
Manual digitizers require the full-time attention of an operator to
position the electronic pen or cursor. The accuracy of various types of
manual digitizers usually is very sensitive to the presence of metallic or
magnetic substances on or near the digitizing surface of the document to
be digitized, to thermal expansion of substances in which grid conductors
of a digitizing surface are embedded, and to the amount of tilt of the
electronic pen. Manual digitizers are quite slow, due to the necessity of
manually moving the cursor. For example, using a typical manual digitizer,
a skilled operator may require approximately eight hours to digitize a
40-inch by 60-inch document, 20 percent of the area of which is covered by
lines that are to be digitized.
Some present automatic digitizers, i.e., those that do not require an
operator to position a cursor or electronic pen for each point or line
digitized, use a video camera to scan a document area within the field of
view of the video camera. Such automatic digitizing devices are extremely
expensive, and their utilization has been limited to applications in which
there has been crucial need for high speed digitizing of entire documents,
such as satellite weather photos, medical X-ray photos, and the like. The
high cost of automatic digitizers has been due to the fact that high
resolution state-of-the-art video cameras are very expensive and a great
deal of high speed data processing capability and memory storage
capability is required to achieve automatic digitizing with high
resolution. The required level of data processing capability and speed has
been roughly equal to that of, for example, a PDP11 minicomputer made by
Digital Equipment Corporation, or equivalent machines. Due to the high
cost of automatic digitizing machines, they have not been developed for
competition in the markets presently dominated by manual digitizers.
Another reason, in addition to their high cost, that automatic digitizers
have not been able to compete in the marketplace with manual digitizers is
that "smart" manual digitizers are readily available at substantially
lower cost. "Smart" digitizers are those that can "interface" with
computers to provide "menu selection" of complex stored shapes (such as
alpha-numerics) that can be almost instantly selected from a "menu" by
appropriate positioning of a cursor or electronic pen and keyboard entry
of commands calling up a selected shape that is stored in digitized form
in the computer, although such stored shapes would be very time-consuming
to digitize by manual positioning of a cursor or electronic pen. The
availability of smart digitizers with this type of menu selection
capability, and the exceedingly high cost of automatic digitizers, has
caused it impractical to attempt to develop low cost automatic digitizing
machines that would compete "head on" in the markets with manual
digitizers. Furthermore, the limited field of view of the state-of-the-art
video cameras and the need for certain minimum levels of resolution has
led to the requirement that video cameras of present automatic digitizing
devices be located relatively close to the document surface, preventing
digitizing of large documents.
However, if automatic digitizers could be constructed inexpensively, the
fact that they are capable of automatic operation without the cost of a
human operator, the fact that they are extremely fast, and the fact that
they are capable of digitizing not only point and line coordinates, but
also degrees of darkness, suggest that automatic digitizers would be
readily accepted if they could be provided inexpensively and with the
capability of providing satisfactory resolution while digitizing large
documents (as large or larger than the sizes that can be digitized by
presently available manual digitizers).
Accordingly, it is an object of this invention to provide an automatic
digitizing machine and method capable of high speed, accurate, high
resolution automatic digitizing of position coordinates of points of large
documents or items, at a cost that is competitive with the costs of
presently known manual digitizers.
It is another object of the invention to provide a digitizing device and
method that allows rapid, accurate digitizing of large documents utilizing
a plurality of video cameras or other array type image sensor devices.
SUMMARY OF THE INVENTION
Briefly described, and in accordance with one embodiment thereof, the
invention provides an automatic digitizer including a plurality of video
cameras within a housing and adjacent to a transparent plate having an
outer digitizing surface and continuous overlapping areas within the
fields of view of the respective video cameras; a removable reference
surface having a plurality of permanent; aligned alignment marks thereon
for placement against the digitizing surface for effecting initializing of
the automatic digitizer to effectuate skew correction and offset
correction of coordinate data produced by each of the video cameras;
circuitry for converting analog video signals produced by the respective
video cameras to corresponding digital numbers representing the relative
darkness of the most recently scanned point of a document or item being
digitized; and circuitry for effecting initial digitizing of the
coordinates of the alignment marks on the reference surface and using
information based on those coordinates to compute skew corrected, offset
corrected document coordinates corresponding to each pair of camera
coordinates produced by each of the video cameras. In the described
embodiment of the invention, multiplexer circuits couple the video outputs
of the respective video cameras to the input of a video amplifier and to a
cursor injection circuit. The output of the cursor injection circuit is
coupled to a video monitor that displays the portion of a document that is
within the field of view of the presently selected camera.
A cursor image is displayed on the video monitor screen and is superimposed
on the portion of the document displayed on the screen. The position of
the cursor (in one embodiment of the invention) is determined by the
position of a joystick. Circuitry responsive to the position of the
joystick produces digital signals representing the position of the
joystick to an input of the cursor injection circuit. Digital information
produced by the presently selected camera is used to increment a
horizontal position counter and a vertical position counter, the outputs
of which are coupled to respective ones of a first set of compare inputs
of a digital comparator. The corresponding ones of a second set of
comparison inputs of the digital comparator are connected to the
respective terminals of an output port of a microprocessor system that
produces, at that output port, logic levels representing the coordinates
of the point to be digitized next. When the contents of the horizontal and
vertical position counter match the coordinates of the next point to be
digitized, the digital comparator produces a convert command. An
analog-to-digital converter has its analog input coupled to receive the
video signal being produced by the presently selected camera, and converts
the amplitude of that video signal to a digital representation of the
relative darkness of the selected point of the document in response to the
convert command.
The microprocessor system operates on that digitized number, which is a
camera coordinate of the field of view of the presently selected camera,
to produce a document coordinate that is corrected for the skew and offset
between the coordinate axes of the field of view of the selected camera
and a document being digitized. To this end, the microprocessor system
executes an initialization process wherein for each camera, the alignment
marks on the alignment surface are digitized. From this information,
correction constants are computed for that camera for the alignment points
in the field of view of that camera. The skew corrected coordinates of the
alignment points are used to obtain skew and offset corrected constants
that are algebraically added to the skew corrected camera coordinates to
produce skew corrected, offset corrected document coordinates that are
uniformly spaced and aligned along transitions from the field of view of
one camera to the field of view of an adjacent camera.
In the described embodiment of the invention, the microprocessor performs a
"dark check" operation to determine the relative light level under "no
internal light" conditions seen by one of the cameras when the digitizing
surface is completely covered by a lightproof cover. The zero reference
adjustment input of the analog-to-digital converter is repeatedly adjusted
for each of the other cameras until the output of the analog to digital
converter is the same as for the first camera. Zero adjust constants are
stored within the microprocessor system memory and are recalled and used
to set the zero reference input of the analog-to-digital converter each
time the corresponding cameras are respectively selected. The
microprocessor performs a light check operation to obtain and store gain
adjust constants that are recalled and used to set the gain adjust input
of the analog to digital converter each time the corresponding camera is
selected.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a partial cutaway perspective view illustrating the automatic
digitizer of the present invention.
FIG. 2 is a plan view of the bottom surface of the cover of the digitizer
of FIG. 1, showing alignment marks on the bottom surface and overlap
regions of the fields of view of various video cameras contained in the
digitizer of FIG. 1.
FIGS. 3A and 3B in combination constitute a detailed block diagram of the
circuitry contained in the digitizer of FIG. 1.
FIG. 4 is a flow chart illustrating the basic operation of the
microprocessor system contained in the circuits of FIGS. 3A and 3B.
FIGS. 4A-4G constitute a flow chart of firmware executed by the
microprocessor and contained in the circuits of FIGS. 3A and 3B in
accordance with the present invention.
FIGS. 5A-5F are diagrams that are useful in explaining a skew correction
process and an offset correction process performed by the digitizer of
FIG. 1.
FIG. 6 is a schematic diagram of a sync detection, video output and cursor
injection circuit used in the circuit of FIGS. 3A and 3B.
FIG. 7 is a schematic diagram of a video amplifier and inverting buffer
circuit used in the circuit of FIGS. 3A and 3B.
DESCRIPTION OF THE INVENTION
Referring to FIG. 1, automatic digitizer 1 includes a housing 3 that
supports the transparent plate 5, the upper surface of 5'of which is
referred to as the "digitizing surface". In operation, a document to be
digitized is placed face down on digitizing surface 5'. A flexible cover
11 has a white bottom surface 11', shown in FIG. 2, which has a number of
spaced alignment points P1, P2 . . . P20 thereon, as subsequently
explained. Cover 11 can be withdrawn from storage slot 12 and placed so
that surface 11' is directly against digitizing surface 5' to effect an
initialization process that is subsequently described herein, or cover 11
can be placed on top of a document to hold the document flat against
digitizing surface 5'.
A plurality of spaced video cameras 19 arranged in rows are disposed within
housing 3 below digitizing surface 5'. A plurality of "light bars" 17 are
also disposed between rows of video cameras 19 to provide controlled
illumination of the surface to be digitized. Light emitted from light bars
17 passes through transparent plate 5 and digitizing surface 5' to the
surface of the document or item to be digitized, or to the surface 11' on
which the above mentioned alignment marks are permanently disposed.
Automatic digitizer 1 includes a video display unit 13 that shows the
portion of a document being presently scanned by a selected one of video
cameras 19, and also displays a cursor, the position of which is manually
controlled by means of joy stick 9 and/or keyboard 7. Reference numeral 21
designates the location of the electronic circuitry shown in the block
diagram of FIGS. 3A and 3B; reference numeral 23 indicates the location of
suitable power supplies.
Overall understanding of the invention can perhaps be best understood if it
is first understood that only one of video cameras 19 is scanning at a
time, and it scans only that portion of the digitizing surface in its
"field of view". As mentioned above, each of the cameras is rigidly
mounted, and therefor scans a fixed area of the digitizing surface 5'. For
the presently described embodiment of the invention, there are 12 video
cameras 19. They are rigidly attached along spaced intervals so that their
respective fields of view include 12 overlapping areas of digitizing
surface 5'. Assume that surface 11' cover 11 is placed on digitizing
surface 5'. The array of video cameras then "sees" surface 11', as shown
in FIG. 2, whereon 20 of the above mentioned alignment marks P1, P2 . . .
P20 are permanently marked. The fields of view of the 12 video cameras 19
are indicated in FIG. 2 by the 12 overlapping solid rectangles, such as
14-1, 14-2 . . . 14-12. The 12 areas bounded by the dotted lines represent
the bounds of skew-corrected, offset corrected x and y coordinate
variables obtained from the corresponding 12 video cameras 19.
It should be understood that the positioning of each of video cameras 19 is
not exact. Therefore, the x and y scanning coordinate axes of each video
camera are skewed relative to horizontal and vertical coordinates of
surface 11.
Therefore, the automatic digitizer 1 causes each x and y "camera
coordinate" digitized for each of video cameras 19 to be "skew corrected".
Each of the 12 video cameras except the one whose field of view includes
the area designated "Camera #1" in FIG. 2 is horizontally and/or
vertically offset from Camera #1. (Hereafter, the 12 video cameras will be
individually referred to as Camera No. 1, Camera No. 2, etc., in
accordance with the field of view area shown in FIG. 2). Therefore, for
each camera except Camera No. 1, each x or y camera coordinate digitized
must also be "offset corrected" relative to Camera No. 1. The basic skew
and offset correction operations will be described next.
To this end, FIG. 5A shows a generalized portion of surface 11' that is
within the field of view 14, hereinafter referred to as the "skewed camera
plane", of one of video cameras 19. Points A and B in FIG. 5A can be any
pair of alignment marks of surface 11', such as P2 and P7 (FIG. 2), in
field of view 14 of a particular video camera.
Since points A and B are accurately marked on the test pattern surface of
the underside of cover 11, the length of line AB in FIG. 5A is a known
constant. Assuming a camera skew condition as shown in FIG. 5A, the
camera, starting at the upper left corner of its field of view 14, scans
alignment point A at camera "alignment point coordinates" x.sub.A,
y.sub.A, and scans point B at camera coordinates x.sub.B, y.sub.B. It can
be easily seen that the video camera having field of view 14 (also
referred to as skewed camera plane 14) is skewed relative to the
horizontal and vertical coordinates of surface 11', which coordinates are
referred to hereinafter as "document coordinates".
As mentioned above, in FIG. 5A, reference number 14 designates the skewed
document plane, as seen by the presently selected camera, which is assumed
to be mounted so that its field of view is slightly skewed. Reference
numeral 18 refers to the "skew corrected camera plane", which is the field
of view that the selected camera would see if it is rotated so that its
skew is perfectly corrected. Reference numeral 28 designates the "document
plane", which is the document to be digitized. Its placement is arbitrary,
but is shown as aligned with the skew corrected camera plane in FIG. 5A
for convenience, because normally there will be document alignment guides
that facilitate aligning of the document with the skew corrected camera
plane.
FIG. 5B shows camera plane 14 and skew corrected camera plane 18 and
generalized alignment points A and B and a third point C.
By applying basic rules of trigonometry, a right triangle ABC may be
formed. The angle .theta. is the angle between lines AB and BC. Since the
hypotenuse length AB is known, and the length of side AC can be calculated
by subtracting y.sub.A from y.sub.B ; the sine of .theta. is therefore:
sine .theta.=AC/AB
Now the "skew corrected document alignment point coordinates" x'.sub.A,
y'.sub.A values of the alignment point A can be calculated.
The camera "alignment point coordinates" x.sub.A, y.sub.A and x.sub.B,
y.sub.B are shown in FIG. 5B. The skew corrected "alignment point"
coordinates x'.sub.A, y'.sub.B are also shown. A right triangle
geometrically similar to triangle ABC is drawn having a hypotenuse length
equal to x'.sub.A and one side equal to x.sub.A. Therefore, by elementary
trigonometry,
x.sub.A '=(AB/AC)x.sub.A.
Therefore, x'.sub.A =kx.sub.A, where k=AB/AC.
Another right triangle geometrically similar to triangle ABC is drawn in
FIG. 5B having a hypotenuse equal to y'.sub.A and one side equal to
y.sub.A. Therefore,
y'.sub.A =(AB/AC)y.sub.A.
Therefore,
y'.sub.A =ky,
where k=AB/AC.
The relationships of the above equations hold for determining the skew
corrected coordinates of any alignment point, and may be generalized as
x'.sub.n =kx.sub.n (1)
and
y'.sub.n =ky.sub.n (2)
where k=AB/AC for the subject video camera, x.sub.n and y.sub.n are the
camera coordinates of alignment point A, and x'.sub.n and y'.sub.n are the
skew corrected camera coordinates around alignment point A.
It should be emphasized that equations (1) and (2) hold for skew correction
of generalized alignment points around which the skew is to be corrected,
but do not hold true for generalized points on a document to be digitized.
Turning now to FIG. 5C, a generalized document point Pn is shown, in
addition to alignment points A and B. .psi. is the skew angle between the
lines AB and AC. By elementary trigonometry, it can be seen that
sine .psi.=(x.sub.A -x.sub.B)/AB and
cos .psi.=(y.sub.B -y.sub.A)/AB.
Still referring to FIG. 5C, it can be proved that
x.sub.Pn '''=(x.sub.p -x.sub.A) cos .psi.+(y.sub.P -y.sub.A) sine .psi.(3)
and
y.sub.Pn '''=-(x.sub.P -x.sub.A) sine .psi.+(y.sub.P -y.sub.A) cos .psi.(4)
where
x.sub.Pn ''' is the skew corrected distance in the x direction between
document point Pn and alignment point A;
y.sub.Pn ''' is the skew corrected distance in the y direction between
document point Pn and alignment point A;
x.sub.P and y.sub.P are the camera coordinates of the document point Pn;
X.sub.A and y.sub.A are the camera coordinates of alignment point A; and
x.sub.B and y.sub.B are the camera coordinates of alignment point B.
Thus, equations (3) and (4) can be used to obtain the coordinates, relative
to alignment point A, of any generalized document point Pn within the
selected camera's field of view.
FIG. 5D shows the skew corrected camera plane 18 and the above-mentioned
document plane 28 for Camera No. 1. At this time, it is convenient to
identify the specific alignment points P1 and P6 used in FIG. 2, rather
than the generalized alignment points A and B of FIGS. 5A-C. Pn is a
generalized document point in document plane 28. In FIG. 5D, the offset in
the x direction between the edge of the document plane 28 and alignment
point P1 is x'.sub.P1D, and the offset between the left edge of the skew
corrected camera plane 18 and alignment point P1 is x'.sub.P1. Thus, it
can be seen that the offset between the left edge of the document plane 28
and the left edge of the skew corrected camera plane 18 is given by
x'.sub.P1S =x'.sub.P1 -x'.sub.P1D. (5)
Similarly, in the y direction, the offset between the upper edge of
document plane 28 and the upper edge of skew corrected camera plane 18 is
given by
y'.sub.P1S =y'.sub.P1 -y'.sub.P1D. (6)
Equations (5) and (6) allow alignment point P1 to be referenced to the
edges of the document plane 28.
Now that the alignment point P1 has been referenced to the edges of the
document plane 28, it is necessary to reference any generalized point Pn
on the document to be digitized to the edges of document plane 28 by
computing the distance x.sub.Pn "" and y.sub.Pn "" shown in FIG. 5D. From
FIG. 5D, it can be seen that
x.sub.Pn ""=x.sub.Pn '"+x'.sub.P1D (7)
and
y.sub.Pn ""=y.sub.Pn '"+y'.sub.P1D (8)
where
x.sub.Pn '" is the distance in the x direction between document point Pn
and alignment point P1;
y.sub.Pn '" is the distance in the y direction between document point Pn
and alignment point P1;
x'.sub.P1D is the distance in the x direction between the left edge of
document plane 28 and alignment point P1; and
y'.sub.P1D is the distance in the y direction between the upper edge of
document plane 28 and alignment point P1.
Likewise, the location of alignment point P2 shown in FIG. 5E, relative to
the edges of document plane 28, can be found in the same manner used for
point Pn in FIG. 5D, in accordance with equations (7) and (8).
Now that the alignment point P2, which is in the field of view of both
Camera No. 1 and Camera No. 2, has been referenced to the edges of
document plane 28, it is now necessary to reference the coordinates of any
generalized point Pn2 in the field of Camera No. 2 to the edges of
document plane 28, as shown in FIG. 5F.
Referring now to FIG. 5F, X.sub.Pn2 '" is the distance in the x direction
between alignment point P2 and document point Pn2. This distance can be
computed according to equation (3). Similarly, y.sub.Pn2 '" in FIG. 5F can
be computed according to equation (4). This enables us to reference the
coordinates of Pn2 back to the edges of the document plane 28 by computing
x.sub.Pn2 ""=x.sub.P2 ""+x.sub.Pn2 '" (9)
and
y.sub.Pn2 ""=y.sub.P2 ""+y.sub.Pn2 "' (10)
where
x.sub.Pn2 "" is the x coordinate of document point Pn2 referenced to the
left edge of document plane 28;
y.sub.Pn2 "" is the y coordinate of document point Pn2 referenced to the
upper edge of document plane 28;
x.sub.P2 "" and y.sub.p2 "" are the x and y coordinates, computed as
explained above with reference to FIG. 5E, of alignment point P2,
referenced to the edges of document plane 28;
x.sub.Pn2 '" and y.sub.Pn2 '" are the skew corrected coordinates of Pn2
referenced to alignment P2 and computed, as mentioned above, in accordance
with equations (3) and (4).
From the foregoing, it can be shown that for a generalized point Pnj in the
field of view of the jth camera, the skew corrected, offset corrected
coordinates referenced to the edges of the document plane 28, are
x.sub.Pnj ""=x.sub.Pj ""+x.sub.Pj "' (11)
and
y.sub.Pnj ""=y.sub.Pj ""+y.sub.Pj "' (12)
where
x.sub.Pj "" is the skew corrected, offset corrected x coordinate,
referenced to the left edge of document plane 28, of the upper left
alignment point in the field of view of the jth camera computed in
accordance with equation (7);
x.sub.Pnj '" is the skew corrected distance in the x direction between the
document point Pnj and the upper left alignment point in the field of view
of the jth camera, computed in accordance with equation (3);
y.sub.Pj "" is the skew corrected, offset corrected y coordinate,
referenced to the upper edge of document plane 28 of the upper left
alignment point in the field of view of the jth camera, and is computed in
accordance with equation (8); and
y.sub.Pnj '" is the skew corrected distance in the y direction between the
document point Pnj and the upper left alignment point in the field of view
of the jth camera, computed in accordance with equation (4).
Turning now to FIGS. 3A and 3B, input multiplexer 37 selects which of
cameras 19 is presently being controlled in response to the signal
outputted on camera select bus 87 from output port 85. The input and
output multiplexer circuitry in blocks 37, 39A, and 39B can be implemented
by means of RCA 4051 CMOS integrated circuit multiplexers.
Assuming that the video cameras 19 (FIG. 1) generate "composite" output
signals (which include video brightness data and also horizontal and
vertical synchronization data), it is necessary to separate the horizontal
and vertical "sync" pulses from the video data. Therefore, output 47 of
multiplexer circuit 37 is coupled to inputs of sync detect circuitry 49,
video amplifier circuitry 51, and cursor injection circuitry 53,
subsequently explained. Sync detect circuit 49, which can be implemented
by means of the circuitry shown in FIG. 6, generates the horizontal
synchronization signal on conductor 55 and the vertical synchronization
signal on conductor 57. Sync detection circuit 49 performs the function of
"stripping" the sync pulse information from the composite video signal on
conductor 47. Conductor 57 is connected to inputs of horizontal control
latch circuit 59, page/frame detect circuit 61, and vertical control latch
63. Page/frame detector 61 indicates whether odd or even lines are being
detected from the active camera and "painted" on the CRT monitor 13 of
FIG. 1. Information in horizontal control latch 59 is used to trigger
horizontal counter circuit 67, which begins counting at the beginning of
"painting" of each horizontal line of the monitor screen. Horizontal
counter 67 counts to 1,024 and can be implemented by means of Texas
Instruments 74LS161 counters. Information in vertical control latch 63 is
used to increment vertical counter 71, which also may be implemented by
means of 74LS161 counters. Vertical counter 71 counts to 512. The counter
output information from counters 67 and 71 is fed into digital comparator
circuit 69, which can be implemented by means of Texas Instruments 74LS85
digital comparator integrated circuits. The "real time" count indicating
| | |
