U.S. patent application number 14/483388 was filed with the patent office on 2016-03-17 for iterative calibration ommitting first corrections from a next iteration.
The applicant listed for this patent is HEWLETT-PACKARD INDIGO B.V.. Invention is credited to Dmitry Iofe, Ran Waidman, Tsafrir Yedid Am.
Application Number | 20160078324 14/483388 |
Document ID | / |
Family ID | 55455052 |
Filed Date | 2016-03-17 |
United States Patent
Application |
20160078324 |
Kind Code |
A1 |
Yedid Am; Tsafrir ; et
al. |
March 17, 2016 |
ITERATIVE CALIBRATION OMMITTING FIRST CORRECTIONS FROM A NEXT
ITERATION
Abstract
In an example implementation, a method of iterative calibration
includes, in a first iteration, measuring first values of a
calibration parameter, and determining first corrections for the
measured first values. The method includes, in a next iteration,
measuring next values of the calibration parameter that incorporate
the first corrections. The first corrections are omitted from the
next iteration measurements to provide pseudo iteration
measurements, and the first iteration measurements are averaged
with the pseudo iteration measurements. Second corrections are then
determined based on the averaging.
Inventors: |
Yedid Am; Tsafrir; (Nes
Ziona, IL) ; Iofe; Dmitry; (Nes Ziona, IL) ;
Waidman; Ran; (Nes Ziona, IL) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
HEWLETT-PACKARD INDIGO B.V. |
Amstelveen |
|
NL |
|
|
Family ID: |
55455052 |
Appl. No.: |
14/483388 |
Filed: |
September 11, 2014 |
Current U.S.
Class: |
358/1.15 |
Current CPC
Class: |
G03G 2215/0109 20130101;
G03G 15/00 20130101; G03G 15/5062 20130101; G06K 15/027 20130101;
G03G 2215/0112 20130101; B41J 29/393 20130101 |
International
Class: |
G06K 15/02 20060101
G06K015/02 |
Claims
1. A method of iterative calibration comprising: in a first
iteration, measuring first values of a calibration parameter;
determining first corrections for the measured first values; in a
next iteration, measuring next values of the calibration parameter
that incorporate the first corrections; omitting the first
corrections from the next iteration measurements to provide pseudo
iteration measurements; and averaging the first iteration
measurements with the pseudo iteration measurements; and
determining second corrections based on the averaging.
2. A method as in claim 1, further comprising applying the second
corrections to a subsequent iteration.
3. A method as in claim 1, wherein the calibration parameter
comprises a printing press misregistration parameter and each first
value comprises a misregistration at a particular (x,y) location on
a sheet of print media.
4. A method as in claim 3, wherein determining first corrections
for the measured first values comprises fitting the first values to
a two-dimensional polynomial.
5. A method as in claim 4, wherein the two-dimensional polynomial
comprises the form: misregistration = i = 0 , j = 0 i + j = 2 C ij
X i Y j ##EQU00006## wherein X and Y comprise a coordinate location
on the sheet of print media, and C comprises a correction constant
associated with the coordinate location.
6. A printing device comprising: a print engine to iteratively
print a pattern on a sheet in a color plane; a measurement module
to determine a misregistration of the color plane; an iterative
calibration module to determine a first correction based on
misregistration from a first iteration, apply the first correction
to the print engine for a second iteration, remove the first
correction from a result of the second iteration, average the first
and second iterations to determine a second correction, and apply
the second correction to the print engine for a third
iteration.
7. A printing device as in claim 6, further comprising a measuring
device to measure the misregistration.
8. A printing device as in claim 7, wherein the measuring device
comprises an in line scanner to capture an image of the pattern for
each iteration.
9. A printing device as in claim 6, wherein the pattern comprises
printed points, each printed point having an (x,y) coordinate
location on the sheet.
10. A printing device as in claim 9, wherein the misregistration
comprises an offset between an expected location of each printed
point and a measured location of each printed point.
11. A non-transitory machine-readable storage medium storing
instructions that when executed by a processor of a printing
device, cause the printing device to: take a first color plane
registration (CPR) measurement at a plurality of grid locations on
a media page; determine a first correction constant for each grid
location; adjust the printing device based on the first correction
constants; take a second CPR measurement at the plurality of grid
locations; remove the first correction constants from the second
CPR measurement to determine a pseudo second CPR measurement;
calculate an average of the first CPR measurement and the pseudo
second CPR measurement; and determine a second correction constant
for each grid location based on the average.
12. A medium as in claim 11, wherein determining a first correction
constant comprises fitting the first CPR measurement to a
two-dimensional polynomial.
13. A medium as in claim 12, wherein the two-dimensional polynomial
comprises the form: misregistration = i = 0 , j = 0 i + j = 2 C ij
X i Y j ##EQU00007## and wherein X and Y comprise a coordinate
location on the media page, and C comprises a correction constant
associated with the coordinate location.
14. A medium as in claim 11, wherein determining a second
correction constant comprises fitting the average of the first CPR
measurement and the uncorrected second CPR measurement to a
two-dimensional polynomial.
15. A medium as in claim 14, wherein fitting the average to a
two-dimensional polynomial comprises fitting the average to the
two-dimensional polynomial comprising the form: p n = fit ( i = 1 n
m i - p i - 1 n ) ##EQU00008## wherein p comprises a polynomial
corresponding with measurement m.
16. A method as in claim 1, wherein the averaging further comprises
averaging data from all previous iterations prior to determining
subsequent corrections for a subsequent iteration.
17. A method as in claim 1, further comprising performing a number
of subsequent iterations after said next iteration, each subsequent
iteration comprising omitting corrections from an immediately
previous iteration to obtain measurements for a current iteration
and averaging measurements from previous iterations with the
measurement for the current iteration before determining a set of
new corrections for the next subsequent iteration.
18. A method as in claim 1, wherein the first and next values being
measured are registration values to calibrate registration of a
plane printed by a printing press.
19. A printing device as in claim 6, further comprising the
iterative calibration module to average data from all previous
iterations prior to determining subsequent corrections for a
subsequent iteration.
20. A printing device as in claim 6, further comprising the
iterative calibration module to perform a number of subsequent
iterations after said third iteration, each subsequent iteration
comprising omitting corrections from an immediately previous
iteration to obtain measurements for a current iteration and
averaging measurements from previous iterations with the
measurement for the current iteration before determining a set of
new corrections for the next subsequent iteration.
Description
BACKGROUND
[0001] Calibration is a step often performed both before and during
the use of many devices and systems. Calibration is the process of
determining and adjusting the accuracy of a device to bring it
within the manufacturer's specifications. In some examples,
calibration is a comparison and matching of device parameters
against values that are known to be correct for the device
parameters in question. Such known values can be referred to as the
standard values. In some calibration examples, the device being
calibrated can be compared to a device whose parameters are known
to comply with the standard values for the parameters of the
particular device. Devices and systems that are not properly
calibrated can produce results that are erroneous, of poor quality,
or are substandard in some way relative to the purpose of the
device or system.
BRIEF DESCRIPTION OF THE DRAWINGS
[0002] The present embodiments will now be described, by way of
example, with reference to the accompanying drawings, in which:
[0003] FIG. 1 shows an example of a system that enables iterative
calibration using accumulated data from all iterations;
[0004] FIG. 2 shows a box diagram of a controller suitable for
controlling a print engine to generate printed substrate and
implementing a color plane registration algorithm;
[0005] FIG. 3 shows a box diagram of an example controller suitable
for controlling a print engine to generate printed media/substrate,
and for implementing an iterative calibration process;
[0006] FIG. 4 shows an example pattern printed on a media sheet
that includes examples of measured color plane registration
location data;
[0007] FIGS. 5 and 6 show flow diagrams that illustrate example
methods related to an iterative calibration process that uses
accumulated data from previous iterations to achieve an accelerated
calibration convergence.
[0008] Throughout the drawings, identical reference numbers
designate similar, but not necessarily identical, elements.
DETAILED DESCRIPTION
[0009] As noted above, calibrating a device helps to adjust the
device performance to bring it within the manufacturer's
specifications. Properly calibrated devices can produce results
that are more accurate and/or of higher quality than devices that
have not been properly calibrated. Printing devices are examples of
devices that can be calibrated to help improve the quality of
printed output. In some printing devices, calibration can involve
matching colors on a monitor with colors on printed output,
optimizing ink distribution on the print media, providing a uniform
progression of ink tints without tonal distortions, and so on. In
addition, because different print media can absorb ink differently,
a separate calibration can be used for each media type.
[0010] One example of a printing device that can be calibrated is a
digital printing press. Calibrations in digital printing presses
often involve iterative processes that can include color
calibration, color plane registration calibration, front-to-back
calibration, and so on. In an example of an iterative calibration
process, the state of the printing press can be measured during a
first iteration, and parameters can be changed according to the
measurements in order to make corrections. The press state can then
be measured again in a subsequent iteration to determine if the
corrections have brought the press within the manufacturer's
calibration specifications. This iterative process can continue
until the parameters converge to bring the press state into proper
calibration.
[0011] Such iterative calibration processes assume that the press
state is kept constant between iterations, so that measurements
from one iteration are valid for use in subsequent iterations.
However, in some circumstances the press can be noisy and the press
state can change from one iteration to another. Thus, the
correction of one iteration can overshoot the target calibration,
resulting in non-convergence of the calibration parameters and a
calibration failure. In some other circumstances, however, even
when the press is noisy and unstable, the calibration parameters
can converge "accidentally" to indicate a successful calibration.
Unfortunately, such false successful calibrations can occur by mere
chance when a current iteration is executed in the same or similar
press state as the previous iteration. Therefore, the calibration
result can be erroneous, and using the calibration parameters later
on may result in printed output that is unacceptable. In order to
minimize such chance occurrences, the standard deviation (.sigma.)
of the calibration can be limited for successful calibration.
However, this yields more failures for noisy substrates.
[0012] One solution for this problem is to use several prints
(i.e., to print several pages) for each iteration. This solution
assumes that different types of substrates (i.e., media sheets)
will experience different press states, and that an average of the
measurements can then be taken over an ensemble of press states.
However, it turns out that when pages are reproduced during
printing, the press state can change even more dramatically than it
changes between sequentially printed sheets.
[0013] Accordingly, example methods and systems described herein
enable an iterative calibration process that uses accumulated data
from previous iterations to achieve an accelerated calibration
convergence. The described method involves accumulating data from
all previous calibration iterations while treating the data as if
all iterations were performed at the same system state (e.g., all
iterations printed at the same press state). In general, data
measured from all iterations is averaged together to make
corrections for a next iteration. However, prior to averaging the
data, corrections previously applied are removed from data measured
in the current iteration. Because the method collects data from a
broad group of system states, the calibration is more robust. The
method improves calibration results, accelerates calibration
convergence, and allows convergence in circumstances where a system
is noisy (e.g., a noisy press), and where prior calibration methods
fail to converge.
[0014] In one example, a method of iterative calibration includes,
in a first iteration, measuring first values of a calibration
parameter, and determining first corrections for the measured first
values. The method includes, in a next iteration, measuring next
values of the calibration parameter that incorporate the first
corrections. The first corrections are omitted from the next
iteration measurements to provide pseudo iteration measurements,
and the first iteration measurements are averaged with the pseudo
iteration measurements. Second corrections are then determined
based on the averaging.
[0015] In another example, a printing device includes a print
engine to iteratively print a pattern of dots in a color plane, and
a measurement module to determine a misregistration for each
printed dot. The printing device includes an iterative calibration
module to determine a first correction based on misregistrations
from a first iteration, apply the first correction to the print
engine for a second iteration, remove the first correction from a
result of the second iteration, average the first and second
iterations to determine a second correction, and apply the second
correction to the print engine for a third iteration.
[0016] In another example, a non-transitory machine-readable
storage medium stores instructions that when executed by a
processor of a printing device, cause the printing device to take a
first color plane registration (CPR) measurement at a plurality of
grid locations on a media page, and determine a first correction
constant for each grid location. The printing device makes
adjustments based on the first correction constants, and then takes
a second CPR measurement at the plurality of grid locations. The
printing device removes the first correction constants from the
second CPR measurement to determine a pseudo second CPR
measurement, and calculates an average of the first CPR measurement
and the pseudo second CPR measurement. The printing device then
determines a second correction constant for each grid location
based on the average.
[0017] FIG. 1 conceptually illustrates an example of a system 100
that enables calibration convergence using accumulative iterations
as described herein. The system 100 can represent any system or
device that incorporates a calibration process to determine and
adjust parameters in order to bring the system or device into a
calibrated state. As discussed herein below, in one example a
system 100 comprises a printing press.
[0018] System 100 includes a measurement device 103 to measure the
state of the system over a number of iterations. Measuring the
state of the system 100 can include measuring one or multiple
calibration parameters. The system 100 also includes a measurement
module 105 and an iterative calibration module 107. In different
examples, the modules 105 and 107 can comprise hardware,
programming instructions, or a combination of hardware and
programming instructions designed to perform a particular function
or combination of functions. Hardware incorporated into modules 105
and 107 can include, for example, a processor and a memory, while
the programming instructions comprise code stored on the memory and
executable by the processor to perform the designated function.
[0019] A function of the measurement module 105 is to determine an
offset between an expected value of the calibration parameter and
the actual/measured value of the calibration parameter. The
iterative calibration module 107 is to determine a correction based
on the offset from each iteration. Thus, the iterative calibration
module 107 determines a first correction based on an offset
determined in a first iteration, and the correction is applied to
the system 100. The state of the system (i.e., calibration
parameter) is then measured in a second iteration to determine if
the parameter has converged to bring the system into calibration.
When the measurement module 105 determines that an offset remains
between the expected value of the calibration parameter and the
actual/measured value of the calibration parameter, the iterative
calibration module 107 prepares for a next iteration by averaging
the data/measurements from prior iterations. Before averaging the
measurements however, the module 107 removes from the second
iteration measurements, the first correction that was previously
applied from the first iteration. The module 107 then averages the
measurements from the first iteration and the second iteration
(with corrections removed), and determines a next correction based
on the averaged measurements. The next correction is then applied,
and another measurement iteration is performed. Removing the
corrections prior to averaging the two iterations makes it so the
measurements for the two iterations are based on the same
parameters, instead of different parameters. Thus, averaging
measurements from the two iterations after removing prior
corrections enables an appropriate accumulation of data that treats
the all the data as if it were measured at the same system
state.
[0020] FIG. 2 shows an example system 100 implemented as a printing
device 100 that enables calibration convergence using accumulative
iterations as described herein. The printing device 100 comprises a
print-on-demand device, such as a liquid electro-photography (LEP)
printing press. A printing device 100 implemented as an LEP
printing press 100 generally includes a user interface 101 that
enables an operator to manage various aspects of printing, such as
loading and reviewing print jobs, proofing and color matching print
jobs, running a calibration process, handling media substrates, and
so on. The user interface 101 typically includes a touch-sensitive
display screen that allows the operator to interact with
information on the screen, make entries on the screen, and
generally control the press 100. A user interface 101 may also
include other devices such as a key pad, a keyboard, a mouse, and a
joystick, for example.
[0021] An LEP printing press 100 includes a print engine 102 that
receives print media/substrate 104 from one or more media input
mechanisms 106, and outputs printed media/substrate 108 to one or
more media output mechanisms, such as output stacker tray 110.
Print media 104 can be in various forms including cut-sheet paper
104 from a stacked media input mechanism 106 as shown in FIG. 1, or
a media web from a media paper roll input mechanism (not shown). In
general, the print engine 102 generates printed media/substrate 108
in the form of printed jobs and outputs the printed substrate 108
to an output stacker tray 110. When the printed substrate 108 is a
media web, one or more finishing devices may be employed to cut the
printed media web into sheets prior to it being stacked in an
output stacker tray 110. Alternatively, the printed media web may
not be cut into sheets and stacked, but instead may be output to a
media output roll (not shown).
[0022] As shown in FIG. 1, an example LEP printing press 100 also
includes a measurement device 103, such as an in line scanner, to
measure ink patterns printed onto a printed substrate 108. A light
source (not shown) may accompany the measurement device 103 to
provide illumination for reflecting off the printed substrate 108.
In some examples, as described herein below, the measurement device
103 can be used to calibrate the color plane registration of the
printing press 100.
[0023] The print engine 102 includes a photo imaging component, or
photoreceptor 112, sometimes referred to as a photo imaging plate
(PIP). The photoreceptor 112 is mounted on a drum or imaging
cylinder 114, and it defines the outer surface of the imaging
cylinder 114 on which images can be formed. In some examples,
images comprise calibration patterns used in a calibration process.
A charging component such as charge roller 116 generates electrical
charge that flows toward the photoreceptor surface and covers it
with a uniform electrostatic charge. A laser imaging unit 118
exposes image areas on the photoreceptor 112 by dissipating
(neutralizing) the charge in those areas. Exposure of the
photoreceptor 112 creates a `latent image` in the form of an
invisible electrostatic charge pattern that replicates the image to
be printed.
[0024] After the latent/electrostatic image is formed on the
photoreceptor 112, the image is developed by a binary ink
development (BID) roller 122 to form an ink image on the outer
surface of the photoreceptor 112. Each BID roller 122 develops a
single ink color (i.e., a single color separation) of the image,
and each developed color separation corresponds with one image
impression. While four BID rollers 122 are shown, indicating a four
color process (i.e., C, M, Y, and K), other press implementations
may include additional BID rollers 122 corresponding to additional
colors. After a single color separation impression of an image is
developed onto the photoreceptor 112, it is electrically
transferred from the photoreceptor 112 to an image transfer blanket
124, which is electrically charged through an intermediate drum or
transfer cylinder 126. The image transfer blanket 124 overlies, and
is securely attached to, the outer surface of the transfer cylinder
126. The transfer cylinder 126 is configured to heat the blanket
124, which causes the liquid in the ink to evaporate and the solid
particles to partially melt and blend together, forming a hot
adhesive liquid plastic that can be transferred to a print
substrate 104.
[0025] In the case of a printing device 100 that uses a print
substrate 104 comprising cut-sheet paper from a stacked media input
mechanism 106, as shown in FIG. 1, a single color separation
impression of an image is transferred from the image transfer
blanket 124 to a sheet of the print substrate 104 held by an
impression cylinder 128. The above process of developing image
impressions and transferring them to the sheet of print substrate
104 is then repeated for each color separation of the image. The
sheet of print substrate 104 remains on the impression cylinder 128
until all the color separation impressions (e.g., C, M, Y, and K)
in the image have been transferred to the sheet. After all the
color impressions have been transferred to the sheet of print
substrate 104, the printed substrate 108 sheet comprises the full
image. The printed substrate 108 sheet with the full image is then
transported by various rollers 132 from the impression cylinder 128
to the output mechanism 110.
[0026] In the case of a printing device 100 that uses a print
substrate 104 comprising a media web from a media paper roll input
mechanism 106, the different color separations (e.g., C, M, Y, and
K) of an image are transferred together from the image transfer
blanket 124 to the web of print substrate 104. Thus, the full image
is built up on the blanket 124 prior to being transferred to the
print substrate 104. Here, the imaging process involves
transferring each color separation from the photoreceptor 112 to
the image transfer blanket 124 until all the color separations
making up the full image are present on the transfer blanket 124.
Once all the color separations forming the full image have been
transferred onto the image transfer blanket 124, the inks for all
the color separations are heated on the blanket 124, and the full
image is transferred from the blanket 124 to the web of print
substrate 104. The printed substrate 108 web with the full image is
then transported by various rollers 132 to the output mechanism 110
where it is typically cut and stacked, or rolled onto an output
media roll.
[0027] In a digital LEP printing device 100, images are created
from digital image data that represents words, pages, text and
images that can be created, for example, with electronic layout
and/or desktop publishing programs, cameras, scanners, and so on. A
controller 120 uses the digital image data to control components of
the print engine 102 during the printing process to generate
printed media/substrate 108, such as controlling the laser imaging
unit 118 to selectively expose the photoreceptor 112. Digital image
data is generally formatted as one or more print jobs stored and
executed on controller 120, as further discussed below. In addition
to controlling the printing process, controller 120 controls the
operation of the measuring device 103 and implements an iterative
calibration process to calibrate the color plane registration (CPR)
of the printing device 100 using measurement data accumulated from
all iterations.
[0028] FIG. 3 shows a box diagram of a controller 120 suitable for
controlling a print engine 102 to generate printed media/substrate
108, and for implementing an iterative calibration process to
calibrate the color plane registration (CPR) of the printing device
100 using measurement data accumulated from all calibration
iterations. Controller 120 generally comprises a processor (CPU)
300 and a memory 302, and may additionally include firmware and
other electronics for communicating with and controlling the
components of print engine 102, such as the user interface 101 and
the media input (106) and output (110) mechanisms. Memory 302 can
include both volatile (i.e., RAM) and nonvolatile (e.g., ROM, hard
disk, floppy disk, CD-ROM, etc.) memory components comprising
non-transitory, machine-readable (e.g.,
computer-processor-readable) media that provide for the storage of
machine-readable coded instructions, data structures, program
modules, JDF (job definition format), and other data, generally
executable by a processor 300.
[0029] As noted above, controller 120 uses digital image data to
control the laser imaging unit 118 in the print engine 102 to
selectively expose the photoreceptor 112. More specifically,
controller 120 receives print data 304 from a host system, such as
a computer, and stores the data 304 in memory 302. Data 304
represents, for example, documents or image files to be printed. As
such, data 304 forms one or more print jobs for printing device 100
that each include print job commands and/or command parameters.
Using a print job from data 304, controller 120 controls components
of print engine 102 (e.g., laser imaging unit 118) to form
characters, symbols, and/or other graphics or images on print
media/substrate 104.
[0030] In addition to controlling print engine 102 for printing,
controller 120 also implements an iterative calibration process
with respect to various aspects of the press 100 including color
calibration, color plane registration calibration, front-to-back
calibration, and so on. An example of an iterative calibration
process is described herein below with regard to color plane
registration. In general, color plane registration (CPR) presents
an ongoing challenge within digital printing presses such as
printing press 100. In a multi-color printing process such as
performed by press 100, the image formation process is performed
separately for each of the colors to produce the finished image.
Each image comprises a single color separation referred to as a
"color plane," and the color planes are brought together to form
the finished image. A finished image may not be formed of all the
available colors, but instead may be formed of any one or
combination of the available colors. Where multiple colors are
used, however, the quality of the finished image depends on how
well the color planes are aligned with one another. The alignment
of color planes is referred to as "color plane registration" (CPR),
and images having misregistered (i.e., misaligned) color planes can
appear to lack sharpness and/or be unclear, or have other anomalies
such as a noticeable color shift in the printed color. One of the
causes for CPR error is the changes that the sheet experiences
between separations. While the sheet may shrink locally, it may
broaden out at some other location when the sheet is ironed by the
press drums. Depending on the state of the press 100, sequential
sheets can be ironed differently. In addition to changing between
sequential sheets, the press state can change even more between
print iterations (e.g., between jobs).
[0031] Thus, in one example of an iterative calibration process,
the controller 120 calibrates the CPR of the printing press 100. To
this end, the controller 120 includes a measurement module 105 and
an iterative calibration module 107 stored in memory 302. The
measurement module 105 and iterative calibration module 107
comprise instructions executable on processor 300 to implement the
iterative calibration process that uses a calibration pattern 306
and measurement data accumulated over all calibration iterations to
calibrate the CPR of printing press 100. The iterative calibration
process can be initiated during printing, at a scheduled interval,
and/or upon receiving a user instruction via the user interface
101.
[0032] The iterative calibration module 107 calibrates the CPR by
collecting measured misregistration data over more than one
calibration iteration, and averaging over the measured data
collected from all iterations. In an iteration, a calibration
pattern 306 print job can be printed, and the measuring device 103
(e.g., in line scanner) can measure the printed sheet to provide
the registration of each printed point within the printed pattern.
Thus, each printed point on the sheet has an (x,y) coordinate
location that is measured by measuring device 103. FIG. 4 shows an
example pattern 306 printed on a media sheet 400 that includes
example measured CPR location data 402 from a first iteration,
along with example reference data 404, and data fit 406 to a
two-dimensional (2D) polynomial. Because iterations are printed
using different parameters, to average over all iterations the
misregistration of all printed sheets should be calculated as if
all sheets were printed with the same parameters. Therefore,
measurements are fitted by a 2D polynomial corresponding with the
(x,y) location on the sheet 400. Each iteration is printed with a
different polynomial according to previous iterations. In order to
accumulate data of all iterations together, the next polynomial fit
is calculated by evaluating the misregistration of each (x,y)
location from the misregistration measurements and the polynomial
that was used in each iteration. For example, consider that the
first measurement (m.sub.1) of the misregistration per (x,y)
location is given in the following Table #1:
TABLE-US-00001 TABLE #1 A. X location B. Y location C. HCPR [mm]
[mm] [um] 0 0 25 350 0 50 700 0 37 0 250 -10 350 250 -70 700 250 2
0 500 10 350 500 12 700 500 5
[0033] The plane misregistration can be fitted by a 2D polynomial
of the form
misregistration = i = 0 , j = 0 i + j = 2 C ij X i Y j
##EQU00001##
[0034] Plugging in column C as the misregistration, column A as X,
and column B as Y, the 2D polynomial fit yields corrections (C) as
shown in Table #2:
TABLE-US-00002 TABLE #2 C.sub.20 0.00012 C.sub.11 -0.000049
C.sub.10 -0.06 C.sub.02 0.00079 C.sub.01 -0.43 C.sub.00 34.64
[0035] This polynomial is p.sub.1, and it corresponds to
measurement m.sub.1. The given measurements and the polynomial fit,
per (x,y) location are shown in FIG. 4.
[0036] Consider now that the second iteration yields measurements
(m.sub.2), per (x,y) location as shown in the following Table
#3:
TABLE-US-00003 TABLE #3 X location Y location HCPR [mm] [mm] [um] 0
0 20 350 0 10 700 0 -5 0 250 32 350 250 30 700 250 15 0 500 41 350
500 35 700 500 20
[0037] In order to average over the two iterations and evaluate the
polynomial for the third iteration, the measurements cannot just be
added to one another, because the two iterations were not printed
with the same parameters. The first iteration was printed without
any polynomial (i.e., no corrections), while the second iteration
was printed with a polynomial resulting from the first iteration.
Therefore, the polynomial from the second iteration measurements is
first removed before averaging over the two iterations, in order to
provide the next iteration polynomial, as shown in the following
Table #4:
TABLE-US-00004 TABLE #4 B. evaluation E. average A. 1.sub.st of
1.sub.st polynomial C. 2.sup.nd D. pseudo of 2 itera- measure- per
location measure- 2.sup.nd measure- tions ment (m.sub.1)
(p.sub.1(x, y)) ment (m.sub.2) ment (C - B) (A, D) 25 34.64 20
-14.64 5.18 50 27.89 10 -17.89 16.06 37 49.47 5 -54.47 -8.74 -10
-24.44 32 56.44 23.22 -70 -35.44 30 65.44 -2.28 2 -18.11 15 33.11
17.56 10 14.81 41 26.19 18.10 12 -0.44 35 35.44 23.72 5 12.64 20
7.36 6.18
[0038] The new the 2D polynomial fit (p.sub.2) yields corrections
(C) as shown in the following Table #5:
TABLE-US-00005 TABLE #5 C.sub.20 -0.000018 C.sub.11 0.0000057
C.sub.10 -0.0036 C.sub.02 -0.000044 C.sub.01 0.044 C.sub.00
9.17
[0039] Generalizing the above calculation yields:
p n = fit ( i = 1 n m i - p i - 1 n ) , ##EQU00002##
[0040] where p.sub.0=0.
[0041] FIGS. 5 and 6 show flow diagrams that illustrate example
methods 500 and 600, related to an iterative calibration process
that uses accumulated data from previous iterations to achieve an
accelerated calibration convergence. Methods 500 and 600 are
associated with the examples discussed above with regard to FIGS.
1-4, and details of the operations shown in methods 500 and 600 can
be found in the related discussion of such examples. The operations
of methods 500 and 600 may be embodied as programming instructions
stored on a non-transitory, machine-readable (e.g.,
computer/processor-readable) medium, such as memory 302 as shown in
FIG. 3. In some examples, implementing the operations of methods
500 and 600 can be achieved by a processor, such as a processor 300
of FIG. 3, reading and executing the programming instructions
stored in a memory 302. In some examples, implementing the
operations of methods 500 and 600 can be achieved using an ASIC
(application specific integrated circuit) and/or other hardware
components alone or in combination with programming instructions
executable by processor 300.
[0042] Methods 500 and 600 may include more than one
implementation, and different implementations of methods 500 and
600 may not employ every operation presented in the respective flow
diagrams. Therefore, while the operations of methods 500 and 600
are presented in a particular order within the flow diagrams, the
order of their presentation is not intended to be a limitation as
to the order in which the operations may actually be implemented,
or as to whether all of the operations may be implemented. For
example, one implementation of method 500 might be achieved through
the performance of a number of initial operations, without
performing one or more subsequent operations, while another
implementation of method 500 might be achieved through the
performance of all of the operations.
[0043] Referring now to the flow diagram of FIG. 5, an example
method 500 of iterative calibration begins at block 502, with a
first calibration iteration in which first values of a calibration
parameter are measured. In some examples, the calibration parameter
comprises a printing press misregistration parameter and each first
value comprises a misregistration at a particular (x,y) location on
a sheet of print media. As shown at block 504, first corrections
are determined for the measured first values. In a next iteration,
as shown at block 506, next values of the calibration parameter
that incorporate the first corrections are measured. The method
continues at block 508 with omitting the first corrections from the
next iteration measurements in order to provide pseudo iteration
measurements. That is, the pseudo iteration measurements are the
next iteration measurements with the corrections from the first
iteration removed (i.e., subtracted out). At block 510, the first
iteration measurements are averaged with the pseudo iteration
measurements, and second corrections are determined based on the
averaging as shown at block 512. The second corrections are then
applied to a subsequent iteration, as shown at block 514.
[0044] As shown at block 516, determining the first corrections for
the measured first values (block 504) comprises fitting the first
values to a two-dimensional polynomial. As shown at block 518, in
some examples the two-dimensional polynomial can comprise a
polynomial of the form:
misregistration = i = 0 , j = 0 i + j = 2 C ij X i Y j
##EQU00003##
[0045] wherein X and Y comprise a coordinate location on the sheet
of print media, and C comprises a correction constant associated
with the coordinate location.
[0046] Referring now to the flow diagram of FIG. 6, an example
method 600 related to an iterative calibration process in a
printing device is shown. The method 600 begins at block 602 with
taking a first color plane registration (CPR) measurement at a
plurality of grid locations on a media page. A first correction
constant for each grid location is then determined as shown at
block 604. At block 606, the printing device is adjusted based on
the first correction constants, and a second CPR measurement is
taken at the plurality of grid locations, as shown at block 608.
The first correction constants are removed from the second CPR
measurement to determine a pseudo second CPR measurement, as shown
at block 610. At block 612, an average of the first CPR measurement
and the pseudo second CPR measurement is calculated, and a second
correction constant is determined for each grid location based on
the average, as shown at block 614.
[0047] As shown at block 616, determining the first correction
constant (block 604) comprises fitting the first CPR measurement to
a two-dimensional polynomial of the form:
misregistration = i = 0 , j = 0 i + j = 2 C ij X i Y j
##EQU00004##
[0048] where X and Y comprise a coordinate location on the media
page, and C comprises a correction constant associated with the
coordinate location.
[0049] As shown at block 618, determining a second correction
constant (block 614) comprises fitting the average of the first CPR
measurement and the uncorrected second CPR measurement to a
two-dimensional polynomial. As shown at block 620, in some
examples, fitting the average to a two-dimensional polynomial
comprises fitting the average to the two-dimensional polynomial
comprising the form:
p n = fit ( i = 1 n m i - p i - 1 n ) ##EQU00005##
[0050] where p comprises a polynomial corresponding with
measurement m.
* * * * *