Share Email Print

Proceedings Paper

Chasing the TIRS ghosts: calibrating the Landsat 8 thermal bands
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

The Thermal Infrared Sensor (TIRS) on board Landsat 8 has exhibited a number of anomalous characteristics that have made it difficult to calibrate. These anomalies include differences in the radiometric appearance across the blackbody pre- and post-launch, variations in the cross calibration ratios between detectors that overlap on adjacent arrays (resulting in banding) and bias errors in the absolute calibration that can change spatially/temporally. Several updates to the TIRS calibration procedures were made in the months after launch to attempt to mitigate the impact of these anomalies on flat fielding (cosmetic removal of banding and striping) and mean level bias correction. As a result, banding and striping variations have been reduced but not eliminated and residual bias errors in band 10 should be less than 2 degrees for most targets but can be significantly more in some cases and are often larger in band 11. These corrections have all been essentially ad hoc without understanding or properly accounting for the source of the anomalies, which were, at the time unknown. This paper addresses the procedures that have been undertaken to; better characterize the nature of these anomalies, attempt to identify the source(s) of the anomalies, quantify the phenomenon responsible for them, and develop correction procedures to more effectively remove the impacts on the radiometric products. Our current understanding points to all of the anomalies being the result of internal reflections of energy from outside the target detector’s field-of-view, and often outside the telescope field-of-view, onto the target detector. This paper discusses how various members of the Landsat calibration team discovered the clues that led to how; these “ghosts” were identified, they are now being characterized, and their impact can hopefully eventually be corrected. This includes use of lunar scans to generate initial maps of influence regions, use of long path overlap ratios to explore sources of change and use of variations in bias calculated from truth sites to quantify influences from the surround on absolute bias errors.

Paper Details

Date Published: 2 October 2014
PDF: 20 pages
Proc. SPIE 9218, Earth Observing Systems XIX, 92181A (2 October 2014); doi: 10.1117/12.2063236
Show Author Affiliations
John R. Schott, Rochester Institute of Technology (United States)
Aaron Gerace, Rochester Institute of Technology (United States)
Nina Raqueno, Rochester Institute of Technology (United States)
Emmett Ientilucci, Rochester Institute of Technology (United States)
Rolando Raqueno, Rochester Institute of Technology (United States)
Allen W. Lunsford, Catholic Univ. of America (United States)
NASA Goddard Space Flight Ctr. (United States)

Published in SPIE Proceedings Vol. 9218:
Earth Observing Systems XIX
James J. Butler; Xiaoxiong (Jack) Xiong; Xingfa Gu, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?