Data Set Overview
  =================
    This dataset contains deconvolved High Resolution Visible CCD (HRIV)
    images of the nucleus of comet 103/P Hartley 2.  Clear and color filter
    (350-950 nanometer) images, which were acquired within +/- one hour of
    closest approach on 04 November 2010 at 13:59:47 UTC during the EPOXI
    mission, have been restored to retrieve much of the resolution that was
    lost due to the defocus of the HRI telescope.  Image scales range from
    1.4 to 85.5 meters/pixel.
 
    All images were deconvolved using the Richardson-Lucy method because
    it preserves global photometry as described by Lindler, et al. (2013)
    [LINDLERETAL2013].  This publication along with Lindler, et al. (2007)
    [LINDLERETAL2007] discuss the positive and negative attributes of the
    Richardson-Lucy algorithm and should be read before using the
    deconvolved images in this dataset.
 
    Although the nucleus in many of the images was severely saturated,
    these frames were deconvolved and included in this dataset because
    there may be useful ice grain information (e.g., see Kelley, et al.,
    2013 [KELLEYETAL2013]).  HRIV frames that captured only the grains but
    not the nucleus were also deconvolved and included in this dataset.
 
 
  Data
  ====
    The deconvolved images located in the /DATA/ directory were produced
    from 106 reversibly calibrated, radiance (RADREV) FITS images archived
    in the PDS dataset, DIF-C-HRIV-3/4-EPOXI-HARTLEY2-V1.0.  (See Klaasen,
    et al., 2013 [KLAASENETAL2011] for a description of the RADREV
    calibration.)  The deconvolution process applied the filter-dependent
    point spread functions for the EPOXI stellar target Canopus that are
    archived in the the PDS dataset DIF-CAL-HRIV-6-EPOXI-STELLAR-PSFS-V1.0.
 
    For each RADREV image, results after 25, 50, 100, 200, and 400
    iterations of the Richardson-Lucy algorithm were stored as separate FITS
    files of the following format:
 
      - The primary data unit is the 2-dimensional, deconvolved RADREV
        image in units of Watts/(meter^2*steradian*micron).  Its header
        was extracted the original RADREV file.
 
      - The first image extension is a residual image of the deconvolution
        process.  It provides the residual for each pixel as a number of
        sigma error, which is described by Lindler, et al. (2013)
        [LINDLERETAL2013].  The residual is computed as (input RADREV
        image) minus (restored image reconvolved with the PSF), normalized
        by the estimated error in the pixel.  Pixels with bad data have
        their residual set to a sigma of zero.
 
      - The second image extension is a mask of the pixels used in
        deconvolution.  Saturated data, detected cosmic rays, and bad
        pixels that were ignored are set to zero in this extension.
        A value of one indicates good data.
 
    A preview JPEG image of the 200 iteration result for each of the 106
    HRIV exposures is included in the /DOCUMENT/PREVIEW/ directory.  The
    JPEGs were generated with a square root intensity scale after clipping
    between 0 to 4 Watts/(meter^2*steradian*micron).  The square root scale
    was chosen so that grains could be seen in the images without the
    nucleus.
 
    File Naming Convention
    ----------------------
      Each FITS file in the /DATA/ directory is accompanied by a detached
      PDS label file.  The naming convention for the deconvolved data
      products is HVyymmddhh_eeeeeee_nnn_Diii.ext where:
 
        'HV' identifies the HRIV instrument,
        'yymmddhh' provides the UTC year, month, day, and hour at the
            mid-point of the observation,
        'eeeeeee' is the exposure ID (OBSERVATION_ID in the label),
        'nnn' is the image number (IMAGE_NUMBER in the label) within the
            exposure ID,
        'D' indicates a deconvolved image,
        'iii' gives the number of deconvolution iterations, and
        'ext' is set to FIT for a FITS file or LBL for a label file.
 
    Image Orientation
    -----------------
      A true-sky 'as seen by the observer' 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 (lines) increasing to the top.  This convention
      is specified in the data labels, where the SAMPLE_DISPLAY_DIRECTION
      keyword is set to RIGHT and LINE_DISPLAY_DIRECTION to UP.
 
      The direction to celestial north, ecliptic north, and the Sun is
      provided in data labels by CELESTIAL_NORTH_CLOCK_ANGLE,
      ECLIPTIC_NORTH_CLOCK_ANGLE, and SUN_DIRECTION_CLOCK_ANGLE keywords
      and are measured clockwise from the top of the image when is
      displayed in the correct orientation as defined by
      SAMPLE_DISPLAY_DIRECTION and LINE_DISPLAY_DIRECTION.  Please note
      the aspect of the North celestial pole in an image can be computed
      by adding 90 degrees to the boresight declination given by
      DECLINATION in the data labels.
 
 
  Parameters
  ==========
 
    Imaging Modes
    -------------
      The general properties of the two HRIV modes used for the images in
      this dataset are provided below.  Please note that HRIV images are
      never binned.  For more information, see Hampton, et al. (2005)
      [HAMPTONETAL2005] and Klaasen, et al. (2013) [KLAASENETAL2011].
 
                       X-Size  Y-Size
          Mode Name    (pix)   (pix)   Comments
          ---- ------  ------  ------  ---------------------
            1  FF       1024   1024    Full frame, shuttered
            2  SF1       512    512    Sub-frame, shuttered
 
    Filters
    -------
      The general properties of the eight HRIV filters used for the images
      in this dataset are provided below.  For more information, including
      the effective wavelength for the radiance calibration, see Hampton,
      et al. (2005) [HAMPTONETAL2005] and Klaasen, et al. (2013)
      [KLAASENETAL2011].
 
          Filter         Center Width
          #  Name        (nm)   (nm)   Comments
          -  ----------  -----  -----  ----------------
          1  CLEAR1       650   >700   Not band limited
          2  BLUE         450    100
          3  GREEN        550    100
          4  VIOLET       350    100   Shortpass coating
          5  IR           950    100   Longpass
          7  RED          750    100
          8  NIR          850    100
          9  ORANGE       650    100
 
 
  Ancillary Data
  ==============
    Observational parameters, such as filter, exposure time, range, and
    pixel scale, have been extracted from the PDS data label of the 200
    iteration result for each of the 106 images and stored in a fixed-width
    ASCII table named DECONV_IMAGE_PARAMETERS.TAB, which is located in the
    /DOCUMENT/ directory.
 
    All time-related keywords in the data labels, except
    EARTH_OBSERVER_MID_TIME, are based on the clock on board the flyby
    spacecraft.  EARTH_OBSERVER_MID_TIME provides the UTC when an
    Earth-based observer should have been able to see an event recorded
    by the instrument.
 
    Observational geometry parameters in the data labels and FITS headers
    were extracted from the archived RADREV products.  The parameters
    were computed by the EPOXI data pipeline at the epoch specified by
    the mid-obs UTC, IMAGE_MID_TIME, in the labels.  The exceptions are
    1) the target-to-sun values evaluated at the time light left the
    target that reached the spacecraft at mid-obs time and 2) the
    earth-observer-to-target values evaluated at the time the light that
    left the target, which reached the spacecraft at mid-obs time,
    reached Earth.  Additionally, the parameters were computed using the
    best available SPICE kernels at the time the RADREV products were
    generated; the kernels are specified in the SPICE_FILE_NAME keyword
    in the data labels.  The most recent version of the kernels are
    archived in the PDS SPICE dataset, DIF-C/E/X-SPICE-6-V1.0.
 
 
  Coordinate System
  =================
    Earth Mean Equator and Vernal Equinox of J2000 (EME J2000) is the
    inertial reference system used to specify observational geometry
    parameters in the data labels.
 
 
  Software
  ========
    The observations in this dataset are in standard FITS format with PDS
    labels and can be viewed by a number of PDS-provided and commercial
    programs.  For this reason no special software is provided with this
    dataset.

Confidence Level Overview
  =========================
    The FITS files in this dataset were reviewed internally by the EPOXI
    project.
 
 
  Review
  ======
    This dataset was certified for scientific use on 05 March 2013, during
    a peer review conducted by the PDS Small Bodies Node.
 
 
  Limitations
  ===========
    Lindler, et al. (2013) [LINDLERETAL2013] discuss how to determine
    which iterated images are best for various purposes.  For example, the
    residual image that is included with each deconvolved data product can
    be used to examine the local chi-squared for various regions of an
    image.  Some parts of an image (i.e., smooth regions on the nucleus or
    the jets) are restored after only a few iterations.  More complex
    regions with sharp structures require additional iterations.  However,
    iterating the smooth regions after the local chi-squared statistic has
    reached the value of 1 (one) will continue to amplify the noise, which
    gives an unrealistic texture to the smooth regions of the nucleus.  It
    can also add fine structure to the jets that is not real.
 
    For the grain images, Lindler, et al. (2013) [LINDLERETAL2013] show
    that increasing the number of iterations helps remove the
    non-linearity with source intensity when performing aperture
    photometry.  However, increasing the number of iterations causes
    additional noise amplification and thus can reduce the ability to
    detect weaker sources.  Kelley, et al. (2013) [KELLEYETAL2013] also
    dicusss the use of the deconvolved, grain images.
 
    Several HRIV frames in the dataset were originally compressed on board
    the flyby spacecraft (EPOXI:COMPRESSED_IMAGE_VALUE in the PDS data
    label).  The data provider, D. Lindler, looked at the compressed image
    restorations versus a modified Richardson-Lucy algorithm (see Lindler,
    et al., 2007 [LINDLERETAL2007]) and did not see any significant
    improvement for the compression table used.  Therefore, the same
    unmodified version of the Richardson-Lucy algorithm was used to
    restore all images in this dataset. However, when computing the residual
    images which are normalized by the errors, the compression errors were
    included.