DATA_SET_DESCRIPTION |
Data Set Overview
=================
This data set contains calibrated images of comet 9P/Tempel 1 acquired
by the Deep Impact High Resolution Instrument Visible CCD (HRIV) during
the encounter phase of the mission. These observations were used for
optical and autonomous navigation (NAV) of the flyby spacecraft as
well as for scientific investigations. These data were collected
on 3-4 July 2005.
Software on board the flyby spacecraft used the optical navigation
(OpNav) and autonomous navigation (AutoNav) images to compute the
brightness centroid of the target body for trajectory corrections.
Optical navigation was used for the cruise phase and most of the
encounter phase, until two hours before impact when the AutoNav
system took control of the flyby spacecraft in preparation for
impact imaging. For more information about the AutoNav system,
see Mastrodemos, et. al (2005) [MASTRODEMOSETAL2005].
A NAV observation consisted of one or more packets of data, each
containing an 80-byte header plus a rectangle of image data called a
''snip'', cropped from a full-frame image. Pixels outside the snips
were not returned. The Deep Impact Science Data Center (SDC) at
Cornell University, used the packets for one observation to
reconstruct a raw, full-frame image of 1008 by 1008 pixels with one
common header. The data pipeline flagged pixels outside the snips
as missing but did not capture the locations of the snips within
a frame. The raw, reconstructed images were then input to the
calibration pipeline at the SDC. For more information about the
calibrated NAV images, see the processing section below.
Reduced NAV images were grouped into daily directories by mid-
observation date. A list of the NAV images in the data set is
provided here:
Exposure IDs
Obs Date DOY Minimum Maximum Observation Type
---------- --- ------- ------- ------------------------------
2005-07-03 184 9000000 9000072 Comet imaging
9000105 9000321 Comet imaging
2005-07-04 185 9000332 9000355 Comet imaging
Essential Reading
-----------------
The following documents, located on the Deep Impact Documentation
volume, DIDOC_0001, are essential for the understanding and
interpretation of this data set:
NAV_IMAGES_REPORT.* : Description of how raw and reduced
NAV FITS images were created for and
processed by the Deep Impact science
data pipeline
NAV_HRIV_REDUCED_ENCOUNTER.* : Science-related image indices for
this data set
NAV_HRIV_FILE_NAMES_RAW2CAL.* : Cross-reference of raw and reduced
NAV file names
CALIBRATION_DOC.* : Instrument calibration by Klaasen,
et al. (2006) [KLAASENETAL2006]
INSTRUMENTS_HAMPTON.* : Instrument paper by Hampton, et al.
(2005) [HAMPTONETAL2005]
AUTO_NAVIGATION_MASTRODEMOS.* : Auto-navigation paper by
Mastrodemos, et. al (2005)
[MASTRODEMOSETAL2005]
MISSION_OVERVIEW_AHEARN.* : Mission overview by A'Hearn, et al.
(2005) [AHEARNETAL2005B]
SCLK_CORRELATION.* : Discussion of the discrepancy
between the spacecraft clocks and
UTC
AICD_FLIGHT_HRIV_MRI_ITS.* : Description of the data set and
definitions of label keywords
Related Data Sets
-----------------
The following PDS data sets are related to this one:
DIF-CAL-HRIV-2-NAV-9P-CRUISE-V1.0 : Raw HRIV NAV cruise data
DIF-C-HRIV-2-NAV-9P-ENCOUNTER-V1.0 : Raw HRIV NAV encounter data
DI-C-SPICE-6-V1.0 : SPICE kernels
The related Deep Impact science data sets are:
DIF-CAL-HRIV-2-9P-CRUISE-V1.0 : Raw HRIV cruise data
DIF-C-HRIV-2-9P-ENCOUNTER-V1.0 : Raw HRIV encounter data
DIF-C-HRIV-3/4-9P-ENCOUNTER-V2.0 : Calibrated HRIV encounter data
Processing
==========
The reduced, two-dimensional FITS images in this data set were
generated by the Deep Impact calibration pipeline, maintained by the
project's Science Data Center (SDC) at Cornell University. The NAV
images report, listed above in the recommended reading section,
discusses how the NAV snips were processed into raw images and
describes the steps performed by the calibration pipeline to reduced
the raw data. A summary is provided here.
NAV images were typically acquired during sequences designed to gather
science data. The major difference between a science and NAV image
was that a NAV observation was made of one or more packets of data,
each containing an 80-byte header plus a rectangle of image data
called a ''snip''. Because the snips were cropped from a full-frame
image, not all pixels were returned in many cases. However, some
images were returned in full as a single snip of 1008 by 1008 pixels.
Because a raw NAV observation was received as one or more snips, the
SDC reconstructed a full frame of 1008 by 1008 pixels from the snips
and stored the results as raw a FITS file. These minimally processed
FITS files were archived as the raw NAV data sets in the PDS.
Since the raw NAV data excluded the serial- and parallel-overclock
pixels around the edges of the array, the SDC pre-processed the raw
NAV FITS images to include this information to make the data look
enough like a 1024x1024-pixel, full-frame science image so that the
existing calibration pipeline could be used.
The pipeline performed the following reduction steps to produce the
reduced FITS images in this data set:
- Calibration of temperatures in the FITS header
- Linearization of data number values
- Correction for bias (using pre-determined values from the
BIASVAL1, BIASVAL2, BIASVAL3, and BIASVAL4 entries in the
PROCESSING_HISTORY_TEXT keyword in the PDS labels)
- Subtraction of a dark frame
- Application of a flat field to normalize the data
- Conversion of data numbers to units of radiance for an absolute,
radiometric calibration
The following calibration steps were disabled (however, calibration
files associated with some of these steps, such tables for removing
cross-talk, were included in the calibration subdirectory for
completeness):
- Decompression (NAV data were never compressed)
- Calibration of temperatures in the FITS header
- Removal of electronic cross-talk and smear
- Normalization of quadrant gains (included in flat fields)
- Correction for uneven bit weighting due to analog-to-digital
(ADC) conversion (a unit correction)
- Gap filling
- Removal of random gaussian noise
- Despiking
- Deconvolution
- Geometric calibration
The resulting data were provided in physical units of radiance,
Watts/(meter**2 steradian micron). These data, designated by the
mnemonic ''RADREV'', were not cleaned and are considered reversible
because the calibration steps can be removed to get back to the
original, raw data numbers.
During the calibration process, the pipeline updated the
pixel-by-pixel image quality map, the first FITS extension, so the
following types of pixels could be identified:
- Pixels where the raw value was saturated
- Pixels where the analog-to-digital converter was saturated
- Pixels that were ultra-compressed and thus contain very little
information
- Pixels considered bad as indicated by bad pixel maps
The pipeline also created a second FITS image extension for a
pixel-by-pixel signal-to-noise ratio map.
The calibration steps and files used to reduce each raw image are
listed in the PROCESSING_HISTORY_TEXT keyword in the PDS data label
for that image. For a detailed discussion of the calibration
pipeline and the resulting data, see the instrument calibration
document by Klaasen, et al. (2006) [KLAASENETAL2006].
Applied Coherent Technology Corporation in Herndon, VA, produced the
PDS data labels by extracting parameters from the FITS headers.
Data
====
File Naming Convention
----------------------
The naming convention for the data labels and FITS files was
Hxcccccccccc_eeeeeee_nnn_RR.LBL or FIT, where:
H : HRIV instrument
x : Image usage (A for AutoNav or O for OpNav)
cccccccccc : Spacecraft clock count at the image mid-point
eeeeeee : Exposure ID, same as for science data; image number
within an exposure ID was always 1 of 1
yyyy : Ground-received time (GRT) year
ddd : GRT day of year
hhmmss : GRT hours, minutes, and seconds
nnn : Sequentially increasing image number within
an exposure ID (always 1 of 1 for NAV data)
It is important to note that a different file naming convention
was used for the raw NAV images. A cross-reference of the raw
and calibrated NAV file names is included on the documentation
volume.
FITS CCD Images
---------------
The two-dimensional, CCD images in this data set are in FITS format.
The primary data array contains the image, followed by two image
extensions that are pixel-by-pixel maps which provide additional
information about the primary image:
- The first extension uses one byte of eight, bit flags to
describe the quality of each pixel in the primary image.
The PDS data label defines the purpose of each bit.
- The second extension provides a signal-to-noise ratio for
each pixel in the primary image.
Each image FITS file is accompanied by a detached PDS label. For
more information about the FITS primary image and extensions, refer
to the instrument calibration document.
True-Sky ''As Seen By Observer'' Display
----------------------------------------
A true-sky view is achieved by displaying the image using the
standard FITS convention: the fastest-varying axis (samples)
increasing to the right in the display window and the slowest-
varying axis increasing to the top. This convention is also
defined in the data labels:
SAMPLE_DISPLAY_DIRECTION = 'RIGHT'
LINE_DISPLAY_DIRECTION = 'UP'
The direction to Celestial North and Ecliptic North, measured
clockwise from the top of the displayed image, is provided in PDS
labels by CELESTIAL_NORTH_CLOCK_ANGLE and
SOLAR_NORTH_POLE_CLOCK_ANGLE, both of which assume the correct
display defined by SAMPLE_DISPLAY_DIRECTION and
LINE_DISPLAY_DIRECTION.
Using this convention for Tempel 1 approach images, ecliptic East is
toward the top, ecliptic North is toward the right, and the Sun is
down. After impact, the Flyby spacecraft came out of shield mode
and turned around to lookback at the comet. For lookback images,
ecliptic East is toward the top, ecliptic North is toward the left,
and the Sun is down.
For a comparison of the orientation of HRIV flight images with those
from ground-based calibrations as well as those from the Medium
Resolution Instrument CCD (MRI) and the Impactor Targeting Sensor
CCD (ITS), see the quadrant nomenclature section of the instrument
calibration document.
Parameters
==========
Data Units
----------
Reduced RADREV data are in units of radiance, W/(m**2 steradian
micron). The data are reversible and are not cleaned or deconvolved.
Imaging Modes and Filters
------------------------
The unbinned, full-frame, instrument mode 1 was always used for NAV
images. Also, the only filters used were the two clear filters: 1
or 6, both centered near 650 nanometers. For descriptions of the
imaging modes and filters, please see the Deep Impact instrument
paper by Hampton, et al. (2005) [HAMPTONETAL2005].
During pre-processing, the SDC added overclock rows and columns
around the edges of the raw NAV FITS images so that the data could
be processed by the calibration pipeline used for the science
images. Pixels in the overclock areas were excluded from the
calculation of the values for MINIMUM, MAXIMUM, MEDIAN, and
STANDARD_DEVIATION in the data labels.
Time-Related Keywords
---------------------
All time-related keywords in the data labels, except
EARTH_RECEIVED_TIME, are based on the clock on board the
spacecraft. EARTH_RECEIVED_TIME provides the UTC when an
Earth-based observer should be able to see an event recorded by
the instrument.
Although the time to impact was not calculated for the NAV data,
the estimated time of impact at the flyby spacecraft was UTC
05:44:34.265 on 4 Jul 2005. This is based on the conclusions
presented in the spacecraft clock correlation report provided
in the DI Documentation volume.
Geometry-Related Keywords
-------------------------
The SOLAR_NORTH_POLE_CLOCK_ANGLE in the data labels specified the
the direction of ecliptic north as projected onto the image plane.
It is measured from the 'upward' direction, clockwise to the
direction toward ecliptic north when the image is displayed as
defined by the SAMPLE_DISPLAY_DIRECTION and LINE_DISPLAY_DIRECTION
keywords.
The SDC pipeline was not able to automatically determine the
proper geometric information for the target of choice in many
cases. When these parameters could not be computed, the
corresponding keywords in the PDS data labels were set to a
value of unknown (UNK). Geometry-related keywords for most
calibration targets were set to UNK.
Geometric parameters provided in the data labels were computed at
the epoch specified by MID_IMAGE_TIME, except for the target-to-sun
and earth-observer-to-target parameters. Target-to-sun values were
calculated for the time when the light left the sun while
earth-observer-to-target were calculated for the time when the light
left the target.
Geometry-related parameters in the PDS data labels are uncertain at
a level of a few seconds because of a known 2-second discrepancy
between the clocks on board the flyby and impactor spacecraft and
between in-situ data and ground-based observations. After a
detailed analysis of the timing problem in early 2006, improved
self-consistent SPICE kernels were generated by the Deep Impact
project to correlate the spacecraft clocks; there is still a
1-2 second uncertainty between the in-situ data and the ground-
based observations and an uncertainty of about one half of a
second between the clocks on the flyby and impactor spacecraft.
These improved kernels were included in the DI SPICE data set
and were used to calculate the geometric parameters in the PDS
data labels. For more information about this discrepancy, please
see the spacecraft clock correlation report provided on the DI
documentation volume, DIDOC_0001.
The SPICE kernels used to calculated the geometric parameters are
provided by the SPICE_FILE_NAME keyword in the PDS data labels. The
kernels were listed in the order they were loaded into memory for
processing.
Ancillary Data
==============
Geometric parameters included in the data labels were computed using
the DI SPICE data set archived in the PDS.
Coordinate System
=================
Earth Mean Equator and Vernal Equinox of J2000 (EME J2000) was the
inertial reference system used to specify observational geometry
parameters in the data labels.
|