This program calibrates images from the Messenger mission's MDIS instrument.
The input image is corrected for the following, in order, during
calibration:
The signal level of AMICA is determined by a combination of an electronic offset
which defines the "zero" level of the analog-to-digital converter (the bias) as well
as any additional signal due to thermal electrons (the dark current). In general, the bias
offset depends upon the electronics and CCD temperature. The bias for AMICA was empirically
modeled as a function of time:
BIAS(t) = B0 + B1*t + B2*t2,
where t is in units of days.
The constants
B0,
B1 and
B2 were chosen to
fit the curve to the data as closely as possible with the following values:
Variable |
Value |
B0 |
3.18x10E02 |
B1 |
-4.12x10E-03 |
B2 |
2.00x10E-05 |
The DN level stored in raw images is approximately proportional to the number
of photons detected in each CCD pixel over the exposure time. Linearity is a measure
of the degree to which the CCD response is proportional to the incident flux.
Linearity was tested during the pre-flight and mission phase of the mission. During the
pre-flight phase, DN values taken at different exposure times showed a linear intensity with
an incidence flux between 1000 DN and 3500 DN. By curve-fitting the pre-flight data, the
following equation was obtained:
Ioutput = (Iinput)c +
L0*(Iinput)*exp(L1*Iinput)
Variable |
Value |
Description |
Ioutput |
|
The observed intensity of the input. |
Iinput |
1000 DN - 3500 DN |
The actual intensity of the input. |
c |
1-5.0x10E-08 |
The power index of the output signal. |
L0 |
-4.87x10E-11 |
Linearity correction term determined empirically. |
L1 |
5.09x10E-03 |
Correction for the nonlinearity around the saturation limit
(Iinput > 3800 DN) |
During the mission the CCD was exposed to intense radiation from cosmic rays. This resulted
in some pixels with an aberrant dark current value that was higher than their neighbors
(the hot pixels). The number of hot pixels increased during the mission. This made
accurate measurement of faint objects such as stars difficult. The following hot pixels
were identified and removed from the output images during calibration by setting
their values to ISIS::Null:
Line |
Sample |
Dark Rate (DN/s) |
407 |
300 |
540
|
599 |
408 |
330
|
820 |
14 |
310
|
930 |
624 |
305
|
897 |
716 |
290
|
The AMICA instrument is shuttered electronically. Images are exposed for a certain
exposure time in addition to the vertical charge-transfer period of 12.288 milliseconds.
It is during the charge-transfer period that a read-out smear in the vertical direction
is produced. The read-out smear is predominant in images with exposure times
on the order of 100 microseconds. Originally the mission plan was to correct
for this smear on-board the spacecraft. However, an anomaly in one of the reaction
wheels on October 2, 2005 made this impossible. As a result of this, some of the images
returned by the AMICA instrument are smear-corrected, while others are not. This necessitated
creating a smear model to remove smears from images taken during the descending and ascending
period. The read-out smear brightness
Ismear
for unbinned images is modeled from the observed images as follows:
Ismear(H) = SUM(H=0...Nv -1) { [K*[(Iraw(H,V) -
Isky(H,V)]/Nv] }
Where: K = tVCT/(tVCT+texp)
For binned images a different formula is applied:
Ismear(H) = SUM(H=0...Nv - 1)
{ [ K1*Iraw(H,V) ] }
Where: K1 = (1/Nv)*[tVCT/
(tVCT+texp)]
A correction factor is then applied after the smear component is subtracted
from the input image (I
raw):
Iout = C*(Iraw> - Ismear)
Where: C = 1/( 1 + K1 * ( (B - 1)/2*B ) )
Variable | Description |
Ismear | The read-out smear. |
Iraw | The intensity of the raw data taken with exposure time
texp. |
Isky | The sum of the bias and the dark current (~300 DN). |
tVCT | The vertical charge-transfer period (0.012288 microseconds). |
texp | The exposure time (in microseconds). |
Nv | The number of pixels along the V-direction (1024). |
H | The line number of the pixel. |
L | The sample number of the pixel. |
B | The binning number. |
Performs a correction for pixel-to-pixel variation in CCD response and vignetting
(reduction of image brightness near the periphery compared with the center).
Flat-field images for all bands were acquired using an integrating sphere
at NEC Space Technologies, Ltd. at room temperature (around 30 degrees Celsius).
A flat-field image is one which has constant uniform brightness everywhere. AMICA
was pointed into the integrating sphere to acquire images of a field that is known be
be spatially uniform to an accuracy of approximately 2%. The correction is accomplished
by dividing each pixel of the output image by the corresponding pixel in the
flat-field image.
This step is optional, and the formula used depends on the value of the UNITS user parameter.
If UNITS=RADIANCE, the following formula will be used to convert the raw DN values to radiance
(W/m2/sr/µm):
R = Raw*RadianceStandard*RadianceScaleFactor
Variable | Description |
R | The calibrated radiance (w/m2/sr/µm). |
Raw | The raw DN value. |
RadianceStandard | The standard conversion factor from DNs to radiance. |
RadianceScaleFactor | An adjustment factor for each filter. |
If UNITS = IOF, first the above formula will be used to convert from raw DNs to calibrated Radiance, and then
the following formula will be used to convert the raw DN values to I/F (radiance) units:
IoF = R*[pi*(Rs)2]/F
Variable | Description |
IoF | The calibrated radiance in units of I over F. |
R | The calibrated radiance (w/m2/sr/µm). |
Rs | The distance in Astronomical Units (AU) between the Sun
and the target body.
|
F | The solar flux (w/m2/µm) at 1 AU. |
If UNITS=DN, no output conversion will be performed and the output units will be in raw DNs.
- Any in-flight dark current noise for this mission was buried in the read-out noise (~ 4 DN).
When observing Itokawa, the image intensity had a range of 1000-3000 DN. Therefore,
the dark current was considered negligible, and is not accounted for in this calibration
application (see Ishiguro (2010) ).
- The derivation for the formula for smear-removal in the case of unbinned images
does not appear in any of the references given below. Questions concerning
the derivation steps for this formula should be directed towards the USGS.
- Ishiguro, Masateru, et al. "The Hayabusa Spacecraft Asteroid Multi-band Imaging
Camera (AMICA)".
Icarus 207(2010) 714-731.
- Ishiguro, Masateru. "Scattered light correction of Hayabusa/AMICA data and
quantitative spectral comparisons of Itokawa".
Publ. Astron. Soc. Japan (2014) 66(3), 55 (1-9).
Kris Becker | 2016-01-24 |
Original version
|
Tyler Wilson | 2016-07-08 |
Added calibration correction routines for linearity, PSF, smear removal for binned
and unbinned images, and converting to I/F units. Also added hot pixel removal.
Updated the application documentation. Added routines for converting
a single-band ISIS cube to an OpenCV::OpenMat matrix and vice versa for the purpose
of translating and scaling flat-field images so their dimensions match
some AMICA images which have been cropped. Fixes #3880.
|
Kris Becker | 2016-08-25 |
Determined the LOSSY compression images must be scaled by a factor of
16 to match the LOSSLESS mode for proper calibration. Added the UNIT
option to specify output calibration units as DN, RADIANCE or I/F. This
parameter replaces the IOF parameter. Added option to allow user to
determine to apply PSF correction. Added CONFIG parameter that user
can provide a version of the calibration parameters to apply in amicacal.
|
Tyler Wilson | 2016-11-02 |
Added the "APPLYSMEAR" option which will force the smear correction calibration
to be applied, and a new test for smear correction.
The default is for this option to not be applied except in the case where an image is
being processed for which on-board smear correction was not applied. Also made the
applyPSF boolean parameter into a global variable for consistency with the applySmear
global variable.
|
Kristin Berry | 2017-02-07 |
Removed the "APPLYSMEAR" option as a user-specifiable parameter. The smear correction
will still be applied automatically when an image is being processed for which on-board
smear correction was not applied (NSUB <= 1). Also updated application code for
conformity with ISIS3 Coding Standards. Backward Compatibility Issue: The APPLYSMEAR
parameter was removed. The smear correction will automatically be applied to images with
1 sub-image.
|
Kristin Berry | 2017-02-08 |
Removed the "PSF" correction option, since the PSF correction is not yet working. Related
code has been commented out rather than removed entirely so that this work can potentially
be resumed at a later date. Backward Compatibility Issue: The PSF parameter was removed
since it is not yet working. The amicacal program cannot apply a PSF correction anymore.
|
Jeannie Backer | 2017-11-28 |
Updated radiance calibration to divide by exposure duration, as done with reflectance.
Updated amicaCalibration????.trn file with latest radiance scale factor for zs
(value provided by L. Le Corre, PSI).
Updated to use filter V solar flux on all calculations.
Brought code closer to ISIS3 coding standards. Fixes #5243
|
Lucille Le Corre | 2018-01-30 |
Added DN/S as an output option. Apply dark current current as it was disabled
in earlier versions.
|