Note: Descriptions are shown in the official language in which they were submitted.
~(J4~
This lnvention relates to means and a method for
detecting by the use of a television camera, coded information
on the surfaces of objects moving relative to the viewing axis
of the television camera.
This application is a Divisional of Canadian Application
Serial No. 206,177 filed August 2, 1974.
It is an objec'~ of this invention to provide means
and a method utilizing a television camera to produce in the
camera, an image of a surface carrying coded information, whose
path is arranged to pass through the field of view of the camera,
and to scan in a given direction the image of the object, to
electronically rotate the scan to a more suitable angle, if
necessary, and to analyze the video scan output signal resulting
from such scan.
It is an object of this invention to provide means
and a method utilizing a television camera to scan in a
predetermined direction to detect coded information on a
surface within its field of viewD wherein markings on said sur-
face accompanying said coded information are detected to determine
when the surface is within the field of view of said camera and
to determine the orientation of the information relative to the
scan direction.
It is an object of this invention to provide means
and a method utilizing a television camera to detect said coded
information on a ~urface within its field of view, wherein
markings on said surface accompanying said coded information are
detected to determine the orientation of said surface and
determination o the scan results in electronic rotation of the
television ~can to the extent necessary to achieve a more suitable
angle for scanning the information.
Figure 1 show~ a schemati~ view of parcels bearing
encoded labels in use with the television equipment;
Figure 2 shows a suggested label for use in aeeord
with the invention;
Figures 3 and 4 show schematic views of the output
signal of the television camera scan output during the deteetion
of the presence of the label and the extraction of the information
thereon;
Figure 5 shows the circuitry for deriving information
from the television camera scan output information;
Figure 6 shows a television camera scanning raster;
Figure 6a shows a rotated raster;
Figure 7 shows a raster scanning method in aceord with
the invention;
Figure B shows circuitry for rotating the raster;
Figure 9 (a) and 9 tb) show the horizontal and vertieal
sean signals.
Although the invention eovers the extraetion of eoded
information from the surfaee of an objeet moving relative to the
extraetion means, the most eommon use of the invention is, at -
this time, thought to be, the reading of labels, containing in- -
formation such as destination and contents, in coded form on
pareels. It will of course be realized that, as a result of the
extraetion of sueh information, the pareels may be automatieally
sorted and routed, and inventory and shipping reeords automatieally
eompiled.
A label suitable for use with the preferred embodiment
of the invention is ~hawn in Figure 2.
~he label as sh~wn provides information defining areas
~0 10 ~ordered on two opposite sides by thiek parallel orientation
-- 2 ~
stripes 12, also known herein as location marks between which
the information is arranged so that it may be scanned perpendi-
cular to the parallel lines. In order that ordinary languaye
text may appear on the surface (as shown) without causing
confusion with the coded information, the two stripes 12 are
preferably made a color other than black or dark blue (the
preferred color for any plain language) and the stripes 12 of
a selected lighter color (say red) will contrast with any plain
language writing for the reader. Although provision is thus
made for the writing of plain language, if desired, the reading
of such pla n language forms no part of this invention. When
the encoded information is to be scanned, the surface is
illuminated with a color (here green or cyan) complementary to
the bar coloring, so that the bars, as well as the encoded
information, appear dark in contrast to the background (and hence
render the plain language invisible to the TV camera). The
method of decoding the information involves detecting the
contrast between the coded information and the background. Since
the scan will include not only the label but a portion of the
surface on which the label is placed and a portion of the
conveyor, these portions will preferably contrast with the marks.
However, the logic circuitry for detection of the information
on the label will achieve such detection in almost all cases
whether or not such surface and conveyor contrast with the marks.
If desired, for any reason, and noting the comments regarding
the parcel and conveyor, it will be appreciated that the surface,
coding and illumination could be selected, so that the back-
ground ix dark and the encoded information light. The information,
preferably in binary form, i~ conveyed by bars 10 present or not
?~ i~ a specific location, here in columns separated by the dimensions
hl) ~
SC and r~ws by the dimensions SR. l~e red coloured bars may be
replaced by black in applications where no plain language need
appear. The terminology 'row' and 'column' is selected in
relation to the scan of the image of the label in the television
camera, which will take place (if the label is arranged within
the required angular tolerance) with individual scan lines
transverse to the location bars in Figure 2 with successive scan
lines successively displaced from left to right or from right
to left in the figure.
In the label shown, a binary code is shown, wherein
rows of locations in pairs, disposed from one anothertransversely
relative to the longitudinal extension direction of the bars,
i.e. such as 10 and 14, either have an information bar in one
location or an information bar in the other, except in the start
locations lOS, where two bars appear. Thus a simple parity check
is provided at all locations but the start position. If, in
normal coding position, rows (extending vertically in Figure 2)
having two marks (other than in the start position) or having no
marks, then in accord with well known techniques an error in
encoding may be detected~ Although only 14 coding position (plus
the start positions) are sh~wn, this is for ease of illustration.
It will be obvious that any number of coding positions may be
provided, limited only by the size of the label. Further,
although only one data and one data parity column are shown, it
will be obvious that as many data columns as desired may be used
(preferably combined with a data parity line) limited only by the
width of the label. The rectangular shape of the bars selected
is not essential but is preferable in view of the rectilinear nature
of the television scan detection means, and also demonstrates
fO that the label information may be physically produced by a standard
-- 4 --
lO~h~J ~'
bar printer, printiny the ~utput of a computer.
Figure 1 shows a conveyor 16 with a series of
packages thereon, and it will be noted that these are arranged
at random with regard to the orientation of the stripes 12
relative to the viewing of the camera 18 to be described here-
after. (In certain instances, two camera will be used, focussed
on the same or on adjacent areas of the conveyor path). While
the patent application 206,177 filed August 2, 1974 dealt with
reading only labels oriented within predetermined angular limits
about the viewing axis, this application discloses means
whereby labels oriented up to 360 about the viewing æ is may
be read. Herein, as in the previous application, (limits de-
signed for practical purposes of simplicity of logic design)
are set on the angles of pitch and roll of the label (deviations
of the label from a plane perpendicular to the viewing axis about
axes respectively perpendicular and parallel to the travel
direction of the conveyor). Suitable limits in this regard are
+15 for pitch and +30 for roll.
A television camera 18 is arranged to have a viewing
area on the conveyor indicated by the dotted area 20 and a
viewing axis preferably perpendicular to the plane of the conveyor.
For simplicity the television camera 18 is shown as vertically
disposed over the horizontal conveyor with its viewing axis
disposed vertically theretowards. HoweverO it will in practice
often be found more convenient to locate a 45 mirror over the
conveyor to direct the vertical rays from the label and conveyor
at a 90 angle to a horizontally disposed camera. In any event,
the camera is disposed so that the direction of its scan lines
may be measured relative to a datum which bears a predetermined
~0 relation~hip to the conveyor travel direction.
10~
The camera in accord with the invention will be
designed to scan in the conventional raster pattern as indicated
in Figure 6. Figure 6 is, as will be appreciated, simplified
for illustration purposes as there might be something of the
order of 262 lines in each field. (Typicalcameras will provide
interlaced scanning with two fields comprising a complete frame).
In this invention however, the scanning does not provide the
basis for a picture but the results of scanning are analyzed from
the video output signal. Accordingly, if the camera used in the
processes of the invention uses interlaced scanning, each field
may be considered as a separate unit.
In discussing the orientation of the raster, its
direction (indicated by the vector R in figure 6 and perpendicular
rasters by the vectors Rl, R2 in Figure 7) will arbitrarily be
assigned as the direction of the trace lines of the scan, i.e.
those which are sensed as distinct from the retrace lines which
are blanked out in the scan output signal.
It is one of the principal features of the invention
herein described, that the scanning raster of the television
camera may be rotated to scan randomly oriented coded information
at a consistent angle.
Although the means for detecting the random angle which
the coded information assumes in the camera image will be des-
cribed hereafter, the means and method for rotating the tele-
vision scan will now be discussed.
The scan as shown in Figure 6 results as is well known
- from conventional TV camera design from the combination of a saw
tooth signal H for controlling (in conventional cameras) the
horizontal deflection (~uch signal being commonly known as the
horizontal sweep) of the scanning beam with a saw tooth vertical
.. . .. .. .
lO~'~h~ ~
signal. The riyhtward rising portion of the horizontal signal
H (Figure 9a) corresponds to the scanning trace when the camera
image scanned thereby is reproduced in the video scan output
signal.
The rightward, sharply falling portion of the saw tooth
(Figure 9a) represents the retrace position (dotted in Figure 6)
which is blacked out in the television scan output signal.
Although the vertical deflection signal V is also a
saw tooth, it represents, during a single field, a signal,
uniformly increasing with time and is so portrayed on Figure 9~b).
The deflection forces controlling the direction of the
scan during the duration of a single scan line are therefore
a horizontal vector H uniformly increasing with time and the
vertical vector V uniformly increasing with time and producing
the resultant vector R, which also represents the raster direction
of Figure 6~
The signals represented by the vectors H and V are
provided by the circuitry of the television camera referred to
as Horizontal Sweep and Vertical Sweep signals respectively are
conventially supplied to the Horizontal and Vertical, respectively,
deflection circuits. These connections are altered at least part
of the time, in accord with the invention.
Most elements and circuitry of the television camera
are not disclosed or discussed herein as these are well known and
conventional.
It may be shown that the rotation of the resultant R
or the ra~ter through an angle ~ may be achieved by applying in
lieu of the horizontal and vertical sw~ep signals, the signals
H' = H cos e + KlV sin e
V' = V cos e + K2H sin e
~ Jl'
where H is the horizontal sweep signal supplied by the
conventional camera circuitry.
V is the vertical sweep signal supplied by the
conventional camera circuitry.
H' and V' are the modified signals supplied for
actuation of the Horizontal and Vertical deflection circuitry
respectively.
e is the absolute value of the raster or resultant
rotation angle.
The upper value for the sign of the second term, in
each case represents counterclockwise rotation, while the lower
value in each case, represents clockwise rotation. (It will
be obvious that the conventions could be altered by having
+e represent counterclockwise rotation and -e clockwise and
using only the upper of the two signs in each case -- hawever
the convention set out in the equations above, represents more
closely the preferred circuitry which requires selectively
applied inverters to represent the change of sign.
As stated, the method of detecting the angle of the
information pattern to the scan will now be described.
In Figure 8, about to be discussed the horizontal and
vertical deflection circuits may involve more cOmpGnents than
those shown. Such components are well known (in each case
usually include an amplifier) and are here represented in block
form.
As shown in Figure 8 the Horizontal sweep signal,
instead of being directly applied to the Horizontal deflection
circuitry as in conventional circuits is applied to multiplier
Ml and to Line Ll. The signal on line Ll, by means of a two
position switch i~ alternatively connectible through an invertor
12 or directly to a mwltiplier M~. Multiplication at multiplier
Ml is by cos e and M4 by sin e.
The vertical sweep signal is applied to the input of
multiplier M3 and to line L2. The signal on line L2, by means
of a two-position switch is alternatively connectible through an
invertor 11 or directly to a multiplier M2. Multiplication at
multiplier M3 is by cos ~ and at multiplier M2 by sin e.
The switches to determine each of the invertors -
direct line choices are shown as ganged and for counterclockwise
o e will be in the upper position.
Although these switches are shown schematically,
the switches will preferably be of the electronic, solid state
type in order to provide rapid switching action.
The output of multipliers Ml and M2 is added at adder
Al and the summed output supplied to the horizontal deflection.
The output of multipliers M3 and M4 is added at adder A2 and the
summed output supplied to the vertical deflection circuit.
The above described circuitry is effective to produce
the desired scan rotation through an angle e if the wiring of
the horizontal and vertical deflection means, and the horizontal
and vertical sweep signals are such that (a) the horizontal sweep
signal causes equal vertical deflection when applied either to
the vertical or horizontal deflection circuits and (b) if the
vertical sweep signal likewise has the same deflection effect
whether applied to the vertical or the horizontal deflection
circuit.
It will be obvious rom the circuit, that in this
event, the output of the adder ~1 to the horizontal deflection
circuit will be H cos e + v sin e and the output of the adder
~2 to the vertical deflection circuit will be V cos e + H sin e,
_ 9 _
o ~
the upper and lower signs in each term representiny the upper
and lower positions respectively of the ganged switches and
counterclockwise and clockwise respectively.
It will be noted that most cameras are not equivalently
designed in their horizontal and vertical deflection circuits,
i.e. a signal having a given deflection effect when applied to
the horizontal deflection circuit would have a different
deflection effect when applied to the vertical deflection circuit.
Since the relationship between deflection and signal
amplitude would be linear in each case the difference may be
compensated for by increasing or decreasing the signal by a constant.
Reference is now made to the dotted boxes M5 and M6 in Figure 8.
These represent multipliers by Kl and K2 respectively. Thus K2
represents the constant by which the horizontal sweep signal H
must be multiplied to produce the equivalent vertical deflection
to the horizontal deflection it conventionally would have caused,
while Kl, conversely, represents the constant by which the vertical
sweep signal V must be multiplied to produce the equivalent
horizontal deflection to the vertical deflection it conventionally
would have caused. The vertical arrow to each multiplier M2 and
M6 represents the terminal at which the indicated multiplying
factor is applied.
Thus with the multipliers M5, M6 in the circuit the
equation becomes :
Signal to Horizontal Deflection Circuit
H C08 e - Kl V sin e
Signal to Vertical Deflection Circuit
v co~ e + K2 H sin e
The re~ult is rotation of the raster counterclockwise
~0 for the upper ~ign~ in the term~ and clockwise for the lower signs
- 10 -
b()~;'
through an anyle e.
This is the situation with the preferred form of
camera (the vidicon) for use with the invention. However, it
has been found convenient and is the preferred method to use a
vidicon camera built to provide equal deflection effects in
response to vertical or horizontal signals. In this event of
course Kl and K2 become 1 and the multipliers M5 and M6 need not
be used in the circuit. The preferred circuit therefore does
not use M5 and M6.
In general any television camera may be used although the
equal deflection vidicon is preferred. If intermittent illumination
of the camera is used the camera must of course be of the type
to retain the image from the time of illumination to the time of
scanning.
In the operation of this part of the circuit therefore,
labels randomly disposed are preferably initially scanned in the
normal orientation, thus e is zero degrees and the factors applied
to the multipliers are cos e = 1 and sin e = 0 and the Horizontal
and Vertical sweep signals are H and V as if the novel circuitry
were not included. When, as discussed hereinafter the orientation
of the pattern is determined and hence the necessary rotation
e of the raster in order to scan it the multipliers are then
actuated to cause multiplication by the amounts cos e and sin e
as indicated. The invertors are also switched to create the
desired sign for rotation in the proper sense.
Although little has been said about the flyback or
retrace o~ the horizontal and the vertical saw tooth scan, the
actuation to cause rotation of the raster also causes rotation of
horizontal and vertical retrace in the right direction and of the
right aTnount.
-- 11 --
~V'~h~ ~
Raster rotations of ~reater than ~0 caus~ complications
in the logic required for switching and in fact logic and control
requirements are simpler if raster rotations are maintained at
less than 45. An information pattern is used so that the
information direction is known, so that it may, in fact, be
scanned in either direction. As will be appreciated, being able
to scan from either direction reduces the maximum rotation e to
not more than 90~. The maximum rotation e may be reduced to not
more than 45 by initially scanning the label with two rasters
at 90 to each other, as indicated in Figure 7. The raster may
then be selected which is at less than 45 (or one of the rasters
at 45 if the information scanning direction is located at
exactly 45 to each raster) and rotated through the angle e with
required sense to allow scanning of the information, again as
hereinafter discussed.
The provision of two scans at 90 to each other, may be
achieved either by providing two cameras each initially scanning
at 90 to the other. The angle of each raster to the pattern is
then detected, as hereinafter discussed, and the raster requiring
the smaller amount of rotation is rotated to align the raster with
the desired pattern scanning direction.
As an alternative to the provision of two cameras, a
single camera may be used to scan alternately in rasters at 90 to
each other using the circuitry of Figure 8. Thus on alternate
scans e is made O whereby the signals ~ and V are applied to the
horizontal and vertical deflection amplifiers respectively (cos e -
sin e = o) as previously noted. On the other alternate scans e is
made 90~ so that cos e = O sin e = 1 and the signal H is applied
to the vertical deflection plates while V is applied to the hori-
?~ zontal deflection plates This effects a rotation of the raster
- 12 -
lO~h~l ,'
through g0. It will be appreciated that control of the invertor
may be used to determine the sense of the 90 anyle between the
scans. If the labels may be oriented at any angle in a 360
range and read in either direction the sense of the 90 angle
will not be of major importance. With the rasters scanning
alternatelyO the orientation of the pattern to each raster is
detected, as hereinafter described and a raster requiring 45
or less rotation is then rotated, in accord with the operation of
the circuitry of Figure 8.
In preference to alternating rotation of the scan
signals through alternation of the sine e and cose e multipliers,
however, it is preferred to achieve the initial 90 rotations
by leaving the value of ~ in the circuit of Figure 8 at O and
alternately interchanging the signals applied to the horizontal
and vertical sweep terminals with suitable alteration in saw
tooth frequency and scale.
Although the circuitry of Figure 8 modifying the
conventionally supplied horizontal and vertical sweep signals,
is the preferred method of rotating the scan, it will be under-
stood that the generation of the necessary signals for the rotatedscan may instead (and within the scope of the invention) be
generated within the means for producing the horizontal and
vertical sweep signals themselves, by techniques of signal pro-
duction and shaping well kn~wn to those skilled in the art.
On the basis of the discussion it will be noted that
following the directions of the preferred embodiment without
further modification, not only will the scan of the type indicated
in Figure 6 be rotated, but the rectangular raster shape will also
be rotated. It may be desired to rotate the scan while leaving
the #hape of the pattern unchanged, i.e. a rotation of 90 of
- 13 -
the pattern of ~igure 6 might be desired, while leaviny the
over-all raster rectanyle or 'envelope' oriented as before.
In this event it will be obvious that the trace and retrace
lines must be shorter in travel but greater in number as
illustrated in Figure 6a to produce the raster direction Rl with
the same line density while the vertical signal will have to be
increased in slope to produce the same line spacing. Modifications
of the signal to produce rotations of the raster scan of less
(or more) than 90 while leaving the envelope oriented as
before are more complex but well within well known techniques
of those skilled in the art. This is in addition to the pre-
ferred scan rotation techniques shown and the alternative
techniques discussed it will be understood that modification
of the horizontal and vertical sweep signals either within the
circuitry shown or without is within the scope of the invention.
In general the value H and V of the horizontal and
vertical sweep signals may be altered between their values
before and after scan rotation to produce such dimensional changes
in raster dimensions and raster line density as are required.
Such changes will therefore take place at approximately the
time of rotation of the scan. Thus H and V are not necessarily
constant quantities where circuitry such as that of Figure 8 is
used but may bear different values before and after rotation.
In accord with preferred embodiment of the invention,
the viewing area of the camera or cameras 18 on the conveyor, is
intermittently illuminated by a strobe light 26 (i.e. light
which may be turned on for a short controlled period and then
turned o). In order to allow the use of the red strip 12 as
~riting locations but to have the~e strips contra~t with the back-
ground, the label i~ illuminated with green light so that the
- 14 -
l~J~ J~jS
black marks lO and the red marks 12 both give su~icient
contrast to the camera. It will be appreciated, for the
purpose of decoding-the information, that although it i5 more
convenient to have the information marks within the standard
range of color of a bar printer, it is possible in general,
for both the information and the location marks, to use any
color, which in the illumination provided will contrast with
the background of the label.
The television camera, as is well known, scans the
image formed therein, to provide an electrical current output
(known herein as a 'video scan output signal' or a 'scan output
signal') wherein dark and light areas scanned in each line
produce signals of high and low amplitude. (If desired, equally
available within the known techniques in the art and equally
useful within the scope of the invention, the video scan output
may be provided, for processing by the invention herein des-
cribed, in the form of a larger amplitude signal for the bright
areas scanned and smaller for the dark). As is well known, the
scanning progresses line by line down a field with the video
scan portraying the scanned results of each trace line serially
from the top to the bottom of the field with the scan signals
for each line separated by the retrace which does not appear in
the scan output and is comtemporaneous with the line synchroni-
zation pulses, and so on from one field to the next, with the
fields separated by synchronization pulses known as 'frame sync
pulses'. As is well known, the television camera conventionally
scans, in one field every second line of those required to
completely scan the image, and then scans the omitted lines in the
next field. However, for the purposes of the invention, each
field may be considered as a complete scan of the image, separated
by frame ~ynchron~zation pulses.
lO ~ 3~
The detection of information received through the
line by line scan of a moving image or label places limitations
on how close the information marks may be placed and/or how
fast the label may travel, since the label may travel a material
distance during the scanning of a single frame and will cause
; ambiguity in information marks too closely spaced, having regard
to the speed of the conveyor. Therefore it is preferred to
limit the illumination entering the camera to produce the image,
to a short enough interval that the moving information marks
cannot move sufficiently, or be sufficiently blurred, to be
ambiguous. The most suitable timing for such illumination to
occur, is during the frame sync pulse and the interval of said
illumination is therefore made the approximate interval of
such pulqe. It will be appreciated however, that the length
and occurrence of the short-time illumination may be varied to
suit specific situations and that, once the scan has located
the location marks in the desired position for scanning the
information, the short-time illumination (here sometimes referred
to as 'strobing') takes place at a time relative to the frame
scan, so that the information may be completely scanned between
such strobing.
Thus although the invention applies to the use of a
television camera with deflection and interpretation equipment
as described, with continuous illumination during the location
of the pattern and the reading of the information, it is highly
preferable to use strobed illumination for the reading of the
information and preferable in most applications to use strobed
illumination at all times.
In the logic circuitry reference is made to A~D and
OR gates~ It is assumed however, that for each of such logical
elements, the counterpart inverse loyical element may be
substituted with due attention to the sense of the input and
output signals required for each stage. Thus where an AND
gate is referred to, the gate is of the type where enabling
signals of the same sense are required at all inputs
simultaneously to provide an output of predetermined sense,
the outputs at all other times being of the other senses.
Further, where an OR gate is referred to, the gate is of the type
where an enabling signal of predetermined sense is only
required at at least one 2nput to provide an output of pre-
determined sense, and provides the opposite sense only when
no enabling signal occurs at any of its inputs. Thus by an
AND gate, I include a NAND gate which may be considered a~ an
AND gate with an inverted output, and by an OR gate, I include
a NOR gate which may be considered as an OR gate with an
inverted output. In general the application does not discuss
the relationship between the sense of the output signals of
one stage and the required input to the following stage, it
being realized that it is elementary to those skilled in the
art of logic circuitry that such senses are obviously known
and controllable; and that where the sense at the output of one
stage is the opposite from that required at the next stage,
the necessary inversion may easily be accomplished between
stages.
The video scan output signal of the television
camera (shown in Figure 3 (b)) derived from 'scan A' of
Figure 3(a) is provided to an analogue-to-digital convertor
for the signal. The convertor is designed to discriminate
bet~1een levels in the video scan output signal above and below
a predetermined value. The predetermined value is selected
- 17 -
10~
to be between the level correspondiny to the scan output from
scanning in the illumination provided, a location or
information mark, on the one hand and the level corresponding
to the scan output from scanning the background on the other
hand. The discriminator is designed to provide an output which
has one of two levels, as shown (Figure 3(c)) wherein the two'
levels respectively correspond to video scan output signals
above and below the predetermined level and the Figure 3(c)
level is switched, depending on the crossings of its analogue
input with the predetermined level. The output of the convertor
1 at gate "D" where the 'dark' or information signals are of
high value and the low or background signals are of low value
is applied as one of the inputs to AND gate 4. The convertor
is so designed that a signal which is the inverse output to
that of Figure 3(c) is developed at output 'I' of the convertor
and applied to AND gates 2 and 3, along lines 42 and 44.
A Clock 46 is provided to achieve synchronism in the
logic circuit. ~he clock 46 must pulse at a rate relative to
the television scan rate, and to the dimensions of the in-
formation and location bars so that by sampling the signal ofFigure 3(c) at the leading edge of each clock pulse meaningful
results may be obtained evidencing the spatial relationship
between the location bars and the background, and also between
the information marks and the background. Since the scan rate
is regular and punctuated by line and frame sync pulses, the
number of clock pulses occurring between the start of a line
or other position on the scan line or scan line output and a
spaced location on the same scan line is a measure of distance
along a scan line and a definite width (which it is convenient
to reer to as a 'pulse width') defines the distance travelled
- 18 -
~(3~ 3~
by the scan during the period of the clock pulse, ~ere, as
in the method described, the video scan output is sampled at
the frequency of the clock pulse, it will be obvious that for
the ac~urate extration of information, the length of a
'pulse-width' must be short relative to the dimension in the
scanning direction of the smallest marks to be determined,
namely the information marks.
In practice, the number of pulse-widths (and this
is of course directly related to location and information mark
dimensions in the label design) is preferably 12 for each
location bar and 24 in between. However, for ease of illust-
ration in the drawings, only half the pulse frequency is shown,
i.e. 6 clock pulses during the scanning of each location bar
and 12 between and the specific embodiment is therefore
described using the 6 and 12 clock pulse measures. The location
marks or bars printed by a computer bar printer will have widths
of approximately six pulse widths and a spacing of four pulse
widths in between (three and two respectively in the example).
The rising (here leading) edge of the clock pulse indicated by
transverse lines on the time base (Figure 3 (c) is used to
open gate 4 to sample the output of convertor 1. The results
of such sampling are shown in Figure 3(d). Shown immediately
below in Figure 3 (c) is pulse output fromgate 3 resulting from
the inverted output from gate 1 convertor 1 gated at gate 3
by the output of clock 46.
Por convenience of illustration the finite width of
the clock pulBe iB not shown in the drawings. The clock pulse
lines shown in E'igure 3 therefore represent the leading edge
of the pulse while the negative clock pulse lines of Figure 4a
correspond to the ~railing edge of the clock pulse and in the
-- 19 --
l~J~O~'
preferred embodiment trail the clock pul~e by slightly more
than 1/2 the pulse period~ The state of shift register 5
reflects the state of the inverse signal at gate 2, at ~ampling
times occurring at the frequency of pulses from clock 46 but
out of step therewith as hereinafter described.
Figure 3(a) shows a portion of the image formed
inside the television camera, and scan lines A, B and C following
portions of Figures 3 and 4 are derived from scan line A in
accord with the normal scan of the camera, extending thereacross.
Figure 3 (b) shows the video scan output signal resulting from
scan line A, television camera being conventionally but not
necessarily designed to provide a high amplitude output signal
for dark areas and low amplitude for light areas. The scan
output signal is supplied along the line 01 to the analogue
to digital convertor 1. The convertor 1 is designed, as
previously explained, to discriminate between outputs along line
01 above and below a predetermined level and to provide a
signal of one level when the magnitude is above the predetermined
level and of another level when the magnitude is below the
predetermined level. The predetermined level PL (Figure 3(b)
is selected approximately midway between the magnitude of signal
resulting from the dark information of location marks and the
magnitude of the signal resultin~ from the bright background.
The output of the convertor at terminal D is then shown in
Figure 3(c) as 'digitized video' and is provided along line 40
to AND gate 4. The analogue to digital convertor is also
designed to provide at gate 1 an output which is the inverse of
that shown in Figure 3 to AND gates 2 and 3.
The description of the use of the analogue to digital
3~ convertor in the circuitry of Figure 5 and the illustration of
- 20 -
lU~
Figure 3(b) discuss an analogue to digital convertor wherein
the video scan output signal is compared with the signal level
PL to derive the signal of Figure 3(c) and its inverse. In
applicant's co-pending application serial number 205,673
filed July 25l74, there are described methods of deriving the
signal of Figure 3(c) from the signal itself. The operation
of the circuitry of Figure 5 modified to include means of de-
riving the graph of Figure 3 (c) from the signal itself are
considered within the scope of the invention.
Gate 3 and 4 also have inputs from clock output 46
and are designed to provide an output pulse created by the
leading edge of the clock pulse.
The output of A~D gate 4 (Figure 3(d)) is fed to
counter 7 where the pulse output is counted. The inverse
pulse output of Gate 3 delayed by a convenient fraction of the
clock pulse period to avoid ambiguity with incremented additions
to counter 7 is used to reset counter 7. (Figure 3(e) shows
the gate 3 output without delay). Thus the counter 7 is designed
and connected to count the number of each series of pulses
appearing at the output of gate 4 corresponding to the scanning
of a dark area and to be reset by the first pulse of a series
from gate 3 indicating the beginning of a bright area scanned.
The values in counter 7 are provided to decoder 8 and the decoder
8 is connected to provide decoded outputs when the pulse counts
in counter 7 correspond to the range of widths of a location
bar 12 and within the acceptable height range (which height
determines the bar width in the image). Thus with an expected
width of 6 pulses for a location bar the decoder will be designed
to produce output~ at counts between 5 and 8 inclusiveO When
the counter 7 stand~ at any of these values decoder 8 provides
- 21 -
lO'~h(~ ~
an output on one of the four (i.e. '5', '6', '7', '8') lines
to OR gate 9, produciny at its output an enabling signal to AND
gate 10.
AND gate 10 is also enabled by a pulse from gate 3
(signalling the end of a dark period) along line 46 and from
OR gate 16 when the counter 14 stands at O or '17' - '24'.
Since ~unter 14, as hereinafter explained, is only enabled to
count after a location bar has been scanned, counter 14 is at
O at the beginning of a scan line. Thus starting with scan line
A, as the scan moves from left to right across the frame,
gate 10 provides an output to counter 11, the first time during
the scan of line counter 7 stands at a count of 5-8 at the end
of a dark area.
Thus, in response to the scan crossing location bar
(within the orientation range) or dark area of corresponding
width, counter 11 counts 1 and activates the '1' output of
decoder 12. While the decoder 12 output is '1', AND gate 13,
enabled thereby, provides pulses resulting from the leading edge
clock pulses from clock 46 to counter 14 alo~g line 48 as long
as counter 11 stands at '1'.
The decoder 15 connected to counter 14 provides three
types of output. Firstly, outputs corresponding to counter
values of O and '17' to '24' are connected to OR gate 16 to
provide an enabling signal to gate 10. When counter 14 stands
at these values. The scan length represented by the pulse
counts between 17 and 24 represents the sum of the pulse width
spacing between the location bars (12-16) and the width 5-8 of
the second-scanned location bar, both within the acceptable
range of orientation. Secondly, decoder 15 outputs corresponding
to 1 ~o 16 are provided to gate 17 whose output, in combination
- 22 -
with gate 18, is d~siyned to enable inverted clock pulses
(from clock 46 and inverted by invertor 35) to pasg through
gate 18 when the count on counter 14 is 1-16 inclusive and
to inhibit the passage of such pulses at other times. The
inverted clock pulses are the pulses from clock 46 inverted at
invertor 35 but remaining in synchronism therewith. Thirdly,
output from decoder 15 corresponding to a value of 25 in counter
14 is used to provide a reset signal to the reset terminal 14R
of counter 14 and counter 11.
In operation then with the circuit as described this
far, no signals are provided to the counter 14 until a dark area
(see scan line A) is scanned. If a dark area smaller than 5
pulse widths or larger than 8 pulse widths is scanned, this is
counted on counter 7 but the counter is reset by the first pulse
after the commencement of the pulse of gate 3 at the commencement
of a bright interval and no resultant output occurs at gate 10
since the pulse at gate 3 did not occur when counter 7 stood at
5, 6, 7 or 8. Since there is no output on the decoder 12 '1'
output, counter 14 remains at O and through decoder 15 and OR
gate 17 disables gate 18 so that nothing is shifted into shift
register 15, counter 14 at O also provides one of the three
necessary enabling signals for gate 10.
This state continues until counter 7 has 'counted' a
dark area of between 5 and 9 pulse widths at the time the first
pulse from gate 3 signals the passage by the scan from a dark
to a light area. Then all three inputs to gate 10 are enabled.
The counter 11 then counts '1' indicating that a location bar
(or dar area of similar width) has been scanned. The counter 7
is of course reset after such total count of a dark area by the
dela~ed reset pulse from gate 3.
- 23 -
As soon as counter 11, as descri~ed abo~e, reaches
the count '1'. the output of decoder 12 enables gate 13 and the
resulting clock pulses to pass through gate 13 to counter 14 and
are counted therein from '1' upward causing the output of
decoder 15 for counts from '1' - '16' to disable gate 10 through
gate 16 until at least 17 is reached in counter 14 and to
enable gate 18 through gate 17 for counts from 1-16.
For counts on counter 14 from '1' - '16' the inverted
clock pulse actuates the shift register 5 on the rising
(trailing) side of the inverted pulse and clocks the input
(Figure 4(a)) thereto from gate 2 at intervals trailing the
regular pulse output by the pulse width or approximately 1/2
the clock period. The shift register has 16 positions corres-
ponding to the 16 pulse positions fed thereto during a line
scan. The pattern of pulses produced from the output of gate 2
in shift register is shown in Figure 4(b) where pulses occur
in the areas between the bars, and no pulses occur during
scanning the two information marks 10. Those pulses or their
absence appear as binary signals (pulse or nor pulse) in
successive stages of the shift register. In case it had been
preferable, for the use of the computer, to provide a shift
register 5 carrying record of the presence of pulses during in-
formation marks and no pulses when there are no information marks,
then gate 2 could have been fed from the gate D of the analogue
to digital convertor 1 rather than from the inverse output, and
the contents of the shift register would have~een as shown in
Pigure 4 (c) indicating 2 information bars scanned (scan line A)
between the location bars. In either event the shift register
after clocking by the inverse clock pulses permitted through by
gate 18 contains a ~erie~ of #tage~ contining ~ 'one' or 'zero'
- 24 -
l()~'~h~ ~J
for each pulse position correspondiny to a dark area and a
'zero' or space for each pulse position corresponding to a
bright area or vice versa, and in either event, the record of
the scan in the shift register may easily be read by the
computer. It will be noted that since 16 pulses are read into
the shift register and the space between the bars may be 12-16,
depending on the angle of skew, that the shift register, in
addition to a binary record of the information may have 1-4
stages corresponding to a portion of the second location bar.
However, the location of the stages of the shift register,
corresponding to the second location bar scanned makes the
character of such stages easily detectable by the computer
which will discriminate between an information bar and a
location bar. Note also that the only information row, with
two bars indicates the start of the information so that from
the position of the start bars the computer may detect the
correct order in which the information (which may be scanned
in either orientation) is to be processed.
~s will be obvious, the aceeptable angle of skew and
the consequent tolerance for variation in width of the bar
width as scanned is variable to suit particular design require-
ments and depends upon the accuracy of the scan rotation
performed in accord with the teaching of this invention.
When counter 14 reaches the counts of 17 to 24 inclusive,
the minimum to maximum pulse width of the expected space between
the bars plus the pulse width of the second location bar, has
been scanned. For counts from 17-24 in counter 14, respective
outputs from decoder 15 through OR gate 16 supply an enable
signal to gate 10 which i8 also enabled by the first pulse from
gate 3 ~ignalling the transition from dark to light in the scan.
- 25 -
I f a second dark area of the width of a location bar of 5-
~pulse widths (correct tolerance) is scanned over an interval
ending in counts in counter 14 between 17-24 (correct location
relative to first location bar) then gate lO enabled by
simultaneous enabling outputs at gate 3, 9 and 16 and counter
ll is shifted to the count of 2. The '2' output of decoder 12
is activated to provide one enabling signal to gate 26. Signals
passing gate 26 as hereinafter explained, are counted by
counter l9. Counter l9 is connected to be reset at the time
of the frame sync pulse (i.e. reset between frames) and, when
gate 26 is enabled, counts the number of lines, in a frame,
w~erein the two correctly spaced location bars are detected.
As with the tolerance for the number of placed in shift
register 5 it will be appreciated that the tolerances expressed
above and at various locations in the application, dependent
upon the angle of skew and hence on the tolerance for the
angular rotation of the scan to read the label.
At the sa~e time, as counter 11 moves from 1 to 2,
gate 13 formerly enabled by the 1 output of decoder 12, is
disabled. Counter 14 will be reset at the end of each line,
by a signal derived from the line sync pulse, if it has not
reached 25, or each time it reaches 25, by the '25' output of
its own decoder 15.
When the counter 11 reaches '2' the pattern of in-
formation marks (or any other contrasting material) for the
scan of a single line, will be recorded in shift register 5.
Such pattern will not, however, in the preferred embodiment
of the invention be transferred to the computer until it has
been determined that the whole information label is present in
~0 thefield of view of the television camera and until the direction
- 26 -
1O~
of the scanning raster is such that the info~mation represented
by the information marks or bars scanned in the appropriate
direction (within design limits of skew tolerance) to allo~ the
information desired from the scan to be handled by the computer
or other data treatment means. This is so that the computer
will only receive the line by line information from the shift
register when the position of the label and its orientation
relative to the scan is such that the sequential scan records
from shift register 5 will provide a record of the scan of a
complete label. The determination of the presence of the label
in t~ field of view is achieved by counting, per field, at
counter 19, the number of lines per field in which two properly
spaced location bars are detected. The location of the label
at a desired position in the field of view is also determined
by AND gate 26 provided between the '2' output of decoder 19 ~-and counter 19 to prevent the initiation of counting lines with
two properly spaced location bars by counter 19 until a certain
frame line has been reached. Thus a line counter 22 is
arranged in any desired manner to count the lines of each field
(such as (as shown) by counting line sync pulses and resetting
on every frame sync pulse). A decoder 23 is arranged to provide
; an output corresponding to the desired upper line position
occupied by the label at the time of the frame scan. For
; example, with 262 lines to a field, assuming that a properly
oriented label will encompass 155-160 lines and it may be de-
sirable to detect the label in the upper half of the field, say
between the 40th and 200th lines. The decoder 23 will therefore,
be arranged to provide an output at line counts 40 to 200 (incl.)
to enable gate 26 during this interval to allow the counting of
line~ with properly spaced location bars producing a '2' output
- 27 -
from decoder 12. Counter 19 is connected to be reset by a
signal originated by each frame sync pulse. ~etween such
reset signals counter 19 counts the lines with correctly spaced
double bars starting with line 40. Decoder 20 is designed to
provide an output with the minimum number of good lines f~r
the label being in the correct position has been detected, in
this example 120 lines. The '120' output of decoder 20 is used
to signal the computer, that the shift register 5 will contain
information about a correctly positioned label in the scan
lines of the next field. As separate indication to the computer
will be provided to inform it that any necessary scan rotation
has taken place. If the computer is so programmed, the computer
will on notification of a correctly positioned label and of
the completion of any scan rotation store and process the
successive line recoxds in the shift register resulting from
scanning between the location bars, on the next field. From
the read-out of the shift register the computer may decode
the encoded information.
The use of 120 lines to indicate the presence of a
label whose bars encompass 160 lines is determined by the fact
that it has been found that such determination will ensure that
in substantially all cases the complete label will be in the
next field scanned. Thus the '120' count between lines 40 and
200 may indicate that some lines have not been counted due
to noise in the scan signal or that part of the label is above
line 40 although within the field with 120 lines between lines
40 and 200. In either event the detection of 120 lines will
indicate in a high enough percentage of cases for efficient
operation, that the nèxt field may be used to extract the
information. ~bviously the number 120 will vary with the
- 28 -
h~
illumination, the camera and other parameters. Given the
occurence of a '2' output (or other locations indentification
signal) from decoder 12 or equivalent device, combined with an
ability to count the number of scan lines down a field, there
are many alternative counting or logical arrangements to deduce
the correct positioning of the label. However, such alter-
native arrangements will be dependent on ability to recognize
that two location bars (or other arrangement of location bars)
have been detected in a scan line.
The counters herein are reset as follows :
Counter 7 from delay 6
Counter 11 at each line sync pulse and each count
of '25' from decoder 15
Counter 19 at each frame sync pulse
Counter 14 at each line sync pulse at each count
of '2' from decoder 12
at each count of '25' from decoder 15
Counter 22 at each frame sync pulse
Counter 25 at each frame sync pulse
The feeding of the information to the computer with
a continuously repeating scan of a moving image places limitations
on how close the information marks may be placed and/or how
fast the label may travel, since the label may travel a material
distance during the scanning.
The short interval illumination may be achieved in
various ways. The regular green illumination provided here by
the fluorescent lamps, may be continued while light admitted to
the camera, may be restricted by a mechanical shutter or for
speed, an electro-optical shutter. However, I prefer to provide
a ~trobe light in addition to the fluorescent source, so that,
- 29 -
l~J~h(J ,~
on detection of the label in its correct location, the
fluorescent light may be switched off and the strobe light
turned on and off during the frame sync interval. The strobe
light may be any light source which may be switched on and off
quickly enough to provide the interval within the desired
tolerance and which will provide sufficient illumination to
create a sufficiently bright image for scanning. If red or
other non-black stripes are used it may be necessary to use a
complementary color for the strobe usually by placing a colored
filter in front of a conventional strobe.
The computer will be programmed to detect the output
of decoder 20 and responsive thereto to cause a series of
contents for the shift register to be fed in one of the pulse
forms shown, to the computer. Thus the operation of the
circuitry shown in Figure 5 is the same when the operation is
performed by a continuous scan.
It will be appreciated that the speed and reaction
time of the circuitry and computer software may be sufficiently
fast that there will, in some design alternatives, be the chance
that the same label, scanned to extract the information, be
again detected in the correct location and the information
again scanned. This may be avoided by sufficient spacing of the
labels bearing parcels on the conveyor (which may be assisted
by making the conveyor of the tray-type or of some other divided
type with one label bearing parcel to be placed per division).
Without restriction of the parcel location the logic circuitry
may be augmented to avoid scanning the same parcel, by requiring
that there be detected the absence of the required number of
double bar lines in the scanning range (here between frame
line~ 100 and 160) between one acceptance of information by the
computer and the next.
- 30 -
1{~h(~ i'
l~here will now be described, the mechanism for
detectiny skew or the deflection of the label about an axis
parallel to the viewiny direction of the camera.
In accord with the invention described herein the
skew measurement is used to determine the angle through which
the raster scan must be rotated in order to provide a raster
scan which is parallel (with acceptance tolerance) to the
desired direction for scanning the information.
It will be obvious that the angle of skew may be
determined by counting for two trace lines in a raster of
known spacing the number of clock pulses that occur between a
reference point and a location stripe and vice versa. The
reference point is preferably the beginning or end of a scan
line. The preferred method for determining skew therefore
is (for two spaced scan lines) to count the number of clock
pulses that occur in the time that the scan moves from the end
of the location bar to the end of the scan line. The difference
between the counts for two different lines is a measure of the
'skew' or the angle e between the scan lines and the desired
direction for reading the information (here perpendicular to
the location stripes).
Thus it will be obvious that
Tan e = K.~ Y
~X
Where e is the angle of skew
a Y is the number of lines between the two lines
selected for measurement.
X is the difference in clock counts on the two
lines
K is a constant selected to compensate for any
scale differences between X and Y.
- 31 -
~3~
The ~et~lod of the invention involve~ maintaining
- the number of lines between the two lines selected for
measurement, constant preferably by making the counts whose
difference results in ~ X on the same lines each time.
Thus ~X is proportional to tan e and this value fed
to the computer, programmed for this purpose will allow e to be
determined, and the sign of ~ X will indicate the direction of
the required angular rotation. As will be appreciated the
computer may readily be programmed to provide from the sign
and value of ~ X provided, the settings sin e and cos e for the
multipliers and the setting of the ganged invertor switches to
produce the desired camera rotation.
The preferred method of calculation of the skew of
the label will now be described. The skew angle 5A is shown
in Figure 3. The skew measurement involves the use of a line
counter 22. The line counter is connected to count signals
originating with the line sync pulse and is reset by a signal
originating with the frame sync pulse. Thus, in any frame, the
line counter 22 contains a count indicating the line being
scanned. Two lines are selected sufficiently spaced that a good
skew measurement may be obtained. These lines need not necessarily
be within the information scanning area.
These lines are chosen in the position of the frame
preceding that position where it is desired to scan the infor-
mation for reading. Thus the lines selected might be 102 and 150.
The decoder 23 for line counter 22 is therefore provided with
outputs which enable AND gates 21 and 24 respectively at line
counts 102 and 150 respectively. Each gate 21 and 24 is also
enabled by the '2' output of decoder 12 through gate 26 and by
the clock pulse from clock 46. The output of gate 21 is connected
- 32 -
o iJ
to the 'count-up' directional terminal of a ~i-directional
counter 25. The output of gate 24 i8 connected to the
count-down directional terminal in counter 25. The counter
24 is reset at the end of each frame. Thus, when the 102nd
line is encountered, counter 21 is enabled after the output
of decoder 12 reaches the '2' output, signalling the end of the
second bar. The clock pulses passing gate 21 cause counter 25
to count up and supply at the end of the scan line a measure
of the distance from the second location bar to the scan edge.
The clock pulses are stopped at the end of the 102nd scan
line, by the disabling of the lead from decoder 23. When
the count in counter 22 reaches 150 for the 150th scan line,
gate 24 is enabled and on the enabling of '2' line from the
decoder 12, the pulses are counted down by counter 25 to the
end of the 150th line. The count remaining in the two way
counter after the end of the 150th line is a measure of the
slope or 'skew' of the label i.e. the angle between the then
existent scanning direction 'R' of Figure 6 and the desired
scanning direction; which, with the label shown is perpendicular
to the location stripes. The sign of the count indicates the
sense of the slope, i.e. a positive residual count indicates a
slope as shown in Figure 3, while a negative count will indicate
a slope in the opposite direction. The residual contents of
- the two-way counter 25 after its 'count-up' and 'count-down'
are therefore available for use by the computer, and may, if
desired, be replaced by two separate counters, one forcounting
line 102 from gate 21, the other for counting line 150 from
gate 24. In ~uch alternative the information may be separately
fed to the computer from the counter.
Although the embodiment described suggests the use
~ o~
of the same two raster Lines, each time, to measure sk~w, it
will be noted that the procedure may be modified to provide
that any two raster lines (spaced by a constant number of lines)
may be used once the stripe pattern has been identified.
It will readily be appreciated that the technique
discussed may be adapted to the use of overlation marks of a
different for~ to the stripes shown and parallel or at another
predetermined angle to the desired scanning direction for the
information.
It will readily be appreciated that the methods and
principles above discussed, for determining s~ews may be applied
to each camera of a pair arranged to initially scan at right
angles to each other. Moreover these methods may also be
applied where a single camera only is used and is caused to scan
initially with alternate fields at right angles to each other.
In the latter event the computer will be programmed to apply to
the multipliers values of e = o and e = so for alternate
fields. (The values of 0 and 90 may be replaced by other pairs
of values for e differing by 90). (If the initial values of e,
differing by 90, are not in directions which are an integral
multiple of 90 it will be appreciated that, while well within
the scope of the invention, the necessary programming will some-
what more complicated and inversions will be required where a
rotation of the scan is carried over a direction which is an
integral multiple of 90).
It will be noted that the other limits or parcel
orientations are required so that due to undue deviations from
these orientations the information will not be ambiguous. Thus,
pitch (rotation about an axis parallel to the conveyor but
perpendicular to its motion direction) will tend in the camera
- 34 -
U'i'
image to shorten the in~ormation bars and the space b~tween
them (increasing the apparent skew) while roll, orientation
of the label about the travel direction axis, will tend to
narrow the apparent transverse dimensions and to decrease the
apparent skew. Thickness of the parcel or other article
raising the level of the information surface relative to the
conveyor increases the dimensions of the information bars and
the location bars in the image. All of the suggested limit~
will be determined for the parameters including the mode of
programming the computer, the vertical scan spacing of the
image in the television camera (effectively setting the re-
solution on the vertical dimension) and the clock pulse
frequency, which effectively sets the resolution in the
horizontal dimension.
The preferred embodiment refers to the provision of
scanning with continuous lighting until a label is detected,
; correctly located, followed by the provision of a strobed or
short period illuminated image for scanning the infor~ation.
It will be obvious that, if desired, short interval or strobed
illumination maybe used for both location of the location bars
as well as detection of the information.
Although location bars of specific width and spacing,
and information bars of specific width and spacing are described
in the specific embodiment, it will be obvious that other
arrangements and dimensions of location bars may be used per-
mitting the detection of the location and orientation of such
location bars by suitably designed logic circuitry and that
other shapes or dimensions of information marks may be used with
obviou~ alteration of the logic circuitryO The information
marks ~ill be for binary BystemB~ that iB, the information is
- 35 -
~ 7~
embodied in their presence or absence at specific locations.
The operation of the device in cpnnection with the
scan rotation will n~w be described (it being appreciated that
the means and method for detecting the label presence before
rotation, and the means for reading the label information
after rototion, has already been described).
With the conveyor moving parcels, two scans of the labels
are provided, either under the action of two cameras scanning
at right angles to each other or with a single camera alter-
nately scanning in mutually perpendicular directions. For
simplicity of operation the cameras will be arranged so that
the scanning rasters in all cases will, relative to their
deflection means be at an angle of 0, 90, 180 or 270.
The computer will be programmed to ensure that any
information read from time to time into shift register 5 will
not be read into the computer until after the scan has been
rotated.
With the two scans at right angles to each other,
either simultaneously or alternating, the results of the scan
are handled by the logic circuit (two logic circuits will be
used if there are two cameras) until, as previously described
a label is located in the correct position in the scan image,
as previously described.
With the location of the label in these camera (or
mutually perpendicular scans in one camera) the skew measurements
~X from each camera or each mutually perpendicular scan is fed
to the computer. rrhe computer is then able to determine which
of the scans is at an angle of les~ than 45 to the desired
scanning direction R, (if both scans are at exactly 45 to
the desired ~canning direction the computer will be programmed
to arbitrarily select one of the cameras).
L04~7
The computer will then be programmed to apply the
correct values of sin e and cos e (derived from the skew
measurement), and any necessary switches of the ganged invertor
switch to rotate the scan of the camera whose raster was at
an angle of less than 45 (or the selected camera) to an
orientation substantially parallel with the desired scanning
direction, (here perpendicular to the location stripes). It
will depend on the accuracy with which this rotation can be
performed, whether or not skew corrections will have to be
made thereafter to the results received from shift register 5
after the raster has been rotated. However experience has
shown that scan rotations may be achieved within ~ of the
desired value and that with resultant skew angles of less than
this amount the computer can read the information without the
necessity of these skew corrections.
Although two mutually perpendicular scans are preferred,
it will readily be appreciated that a single scan may be used,
requiring resultant raster rotation of up to 90. Moreover scan
rotations of greater than 90 may be provided with necessary
inversion switching to provide the requisite sine or cosine
sign changes.
When the scan has been rotated, as already described,
the strobe is initiated if it has not been used during the
location stages.
The computer conditioned by the scan rotation then
reads the inforrnation then being stored in the shift register
~ a~ previously described. The logic circuitry inhibiting
reading of the information in shift register 5 until the correct
po~itioning of label i8 determined may not be needed after the
rotation if 'chere i~ no concern about the label having left the
correct area relative to the camera. If there is concern this
logic circuitry may of course be used to disable -the reading of
the shift register 5 until the correct positioning has been
assured.
With the scan rotated and the label correctly positioned
the results of scanning the information bars are sequentially
supplied to the shift register 5 and sequentially read out by
the computer, as previously discussedO A~ shown and discussed
the label is coded here by two bars, to indicate the direction
of the information.
Thus the information may be scanned in either
direction and the correct sequence determined at the computer.
Obviously if more desirable in a particular application, the
computer can be programmed to determine that the scan is from
end to beginning relative to the information and scan in the
opposite direction.
- 38 -
