MULTISPECTRAL THERMAL IMAGER SCIENCE, DATA PRODUCT AND GROUND DATA PROCESSING OVERVIEW. Page: 4 of 5
This article is part of the collection entitled: Office of Scientific & Technical Information Technical Reports and was provided to Digital Library by the UNT Libraries Government Documents Department.
The following text was automatically extracted from the image on this page using optical character recognition software:
products. See  for a description of the science retrievals.
Level 4 products are more advanced analyses that are not
done routinely by the DPAC.
Several of the Level 1 products bear more description. The
first Level 1 product consists of separate images of each band
for which the radiometric calibration has been applied. The
calibration is based on a sophisticated ground calibration
along with an extensive set of on-board sources. The on-
board calibration was lost due to hardware failure after 6
months of science data were acquired, but MTI data products
based on the reflective bands should be nearly as good as pre-
hardware failure. This is because the reflective calibration
was largely using ground calibration data. However, the
thermal calibration is not likely to be as good now as when
the on-board sources were available, particularly for the
longwave infrared bands. This is because the longwave
bands (L, M, N) use HgCdTl detectors, which are
notoriously unstable. The data processing in either case is
The team is using several calibration techniques to make
up for the loss of the on-board calibration sources, in addition
to regular ground-truth campaigns to provide a vicarious
calibration of the system. For example, the satellite is
acquiring regular images of the moon to monitor optical
degradation and perhaps to be used as a calibration source for
the reflective bands. Studies are also underway to determine
if the moon can be used as a thermal calibration source.
Images are being acquired with the detector arrays aligned
parallel with the satellite sweep direction. These are being
done over open water targets that have temperature-
measurement buoys, to monitor the thermal calibration.
Ground targets will be imaged in the same way to monitor the
reflective band calibration.
The co-registration of the spectral bands and the three
sections of the focal plane is performed next at Level 1. The
co-registration is presently done in two ways: all images are
passed through the automated registration and selected
images are co-registered interactively. The automated
registration relies on downlinked satellite position and
orientation information. The position of each pixel on the
focal plane is mapped through the optics, based on a
distortion map measured during ground calibration, and
projected to the surface of the earth. All pixels in all bands
and all SCAs are resampled to a common grid defined in
The interactive registration allows the user to produce
subpixel alignment of bands in the along-track and cross-
track directions. The image is not warped or rotated. The
three segments of the focal plane are joined in a similar way.
Smith et al. describe another co-registration tool used on the
After the bands and three image sections are co-registered,
selected images are georeferenced. The purpose of the
georeferencing process is to align the rows and columns of an
MTI image with the cardinal directions of a map projection.
This is achieved by associating features in a co-registered
MTI image with their corresponding locations on a reference
map. These coordinate pairings are used to solve for the
coefficients of an affine transformation. The affine
transformation is then used to guide a resampling of the MTI
data onto a georegistered output grid. Georeferenced MTI
imagery can be input to a Geographic Information System
(GIS) so that it may be combined with other geospatial
datasets, such as hydrological vectors, rasterized maps, and
ground truth data. Conversely, the locations of features of
interest in an MTI image can be easily extracted once a
georeferenced product has been created.
A few words on user interaction with the DPAC follow.
Access to MTI data is gained through a web interface, after
authorization is granted to the user. The web interface allows
users to find images of interest by providing access to the
DPAC operations database, which contains extensive
metadata for MTI imagery. Once images of interest are
found, users can request imagery over the web interface
(although all requests must be accompanied by signed
paperwork). The data products are subsequently sent to the
user via CD or Digital Linear Tape (web downloads are not
possible at this time). MTI data products are stored in the
Hierarchical Data Format (HDF), with ancillary data
available in the same file as the image data. A header file (in
both text and HDF formats) also accompanies data products.
The HDF-EOS extensions are not used in the DPAC.
IV. THE DPAC PIPELINE
The DPAC automated pipeline is designed to produce all
Level 0 products and most of Level 1. One exception is
georeferencing, which is done at Level 1 and is not presently
automated. Parts of Level 2 are also automated.
The major features of the automated pipeline include: (1)
Automatic processing of data through Level 0 by simply
placing a raw data file in the appropriate location. No user
intervention is necessary for this processing to occur. (2)
Automatic generation of Level 1 and higher processing
requests when Level 0 is complete. (3) Input from the
database to guide the processing, as well as output to the
database to monitor progress and post results. (4) All the
facilities needed to run the data multiple times (e.g., when
new versions of processing algorithms are developed).
The DPAC maintains two databases, one for DPAC (as
opposed to spacecraft) operations information and another to
store spacecraft and payload state-of-health data. The
operations database provides input to the processing
(automated or otherwise), monitors progress of the
Here’s what’s next.
This article can be searched. Note: Results may vary based on the legibility of text within the document.
Tools / Downloads
Get a copy of this page or view the extracted text.
Citing and Sharing
Basic information for referencing this web page. We also provide extended guidance on usage rights, references, copying or embedding.
Reference the current page of this Article.
SZYMANSKI, J.; BALICK, L. & AL, ET. MULTISPECTRAL THERMAL IMAGER SCIENCE, DATA PRODUCT AND GROUND DATA PROCESSING OVERVIEW., article, April 1, 2001; New Mexico. (digital.library.unt.edu/ark:/67531/metadc724581/m1/4/: accessed January 22, 2019), University of North Texas Libraries, Digital Library, digital.library.unt.edu; crediting UNT Libraries Government Documents Department.