[PPS03-P07] Refinement of calibration and conversion from raw data to thermal distribution
~ HEAT: Hayabusa2 Exploration Assistant for TIR ~
Keywords:Hayabusa2, TIR, Calibration, HEAT, Software, Database
2014, and it is scheduled to arrive at the target in summer of 2018. Hayabusa2 has four optical
instruments for scientific observations such as ONC, NIRS, LIDAR, and, TIR. TIR is a thermal
infrared imager whose primary goal is to know natures of the asteroid and to select candidate
landing sites for safe sampling based on thermo-physical properties of the surface. TIR records
infrared radiation from the target as a digital number (DN). The Hayabusa2 TIR science team has
been preparing analysis procedures and application for its observations. A part of this application
was tested in a simulation drill in 2017. Our team has developed a supporting tool called Hayabusa2
Exploration Assistant for TIR (HEAT). HEAT searches TIR data interactively, displays TIR images,
and visualizes thermal models. HEAT is used for visualization, calibration, and analysis.
In this study, calibration is defined as methods of calculating formula using ground test data. TIR
requires conversion from a DN to a surface temperature of the target. HEAT has two calibration
methods. One is a regression using ground test data, and the other is a direct conversion using
interpolation based on near parameter sets from ground test data. These methods are suitable for
interpolation, but incompatible for extrapolation. Therefore, calibration using blackbody model is
HEAT is required to solve three problems. First, a method of calibration using blackbody model has
not implemented. Second, in this project, TIR images need to be converted to temperature one
within three days. However, HEAT takes a long time to calculate a calibration
formula of 328*248 each pixel because calibration uses large numbers of ground test data. Third,
HEAT has not supported a standard format (FITS: Flexible Image Transport System). This is a
standard format used in the astronomy field. Therefore, this study implements an additional fitting
function using blackbody model, refines the calibration and the conversion methods, and establishes
compatibility with FITS.
Conversion of DN to temperature using blackbody model was conducted within the range of the
accuracy of +/- 2 K.
HEAT has been implemented the blackbody model to conduct calibration with high precision. This
makes it possible to extrapolate range of the ground test data.
HEAT can reduce the calibration and the conversion time by two improvements. One is an
enhancement of processing capability by parallel processing. The other is a reduction of computation
quantity by masking TIR image. These improvements in HEAT make to convert TIR images to
temperature ones within the short period, three days.
To make a calibration formula based on imaging conditions before the conversion is a further
improvement in this processing.
This study refined the calibration and the conversion of HEAT enough to be used in the Hayabusa2
mission. HEAT will be used by TIR science team at the rendezvous phase. HEAT has been
developed to satisfy all requirements.