nodel is ob- 
a continuous 
by 
Grauwert 
9 
( ) 80 4 
60 1 
| 40 | 
20 - 
fi x 
0 ; ; ; E ; Zeile 
0 80 160 240 320 400 485 560 
  
  
Fig. 10: Definition of window for Fig. 11: Mean values of lines 
Zeile regression 
  
  
  
  
  
  
Grauwert 
——— Spalte 
80 560 
80 1 
(upper) 
60 den eth AS ANA AA 
us image 40 
mean 201 
0 t T T T T T Zeile 
0 80 160 240 320 400 480 560 
Fig. 12: Result after SRF Fig. 13: Profile of row after SRF 
ined by the 
RESULTS AND DISCUSSION 
The accuracy of the results is controlled by a comparison with ground-truth boat 
(10) measurements of the temperature with radiation and bucket thermometer readings. 
An increase of accuracy can be seen in table 1 for the use of laboratory calibra- 
tion and atmospheric corrections especially with a SRF-prefiltering. The data used 
hows the for laboratory calibration were obtained more than one year after image recording. 
di Scan- Precise and simultaneous laboratory calibration data seem to be very necessary for 
à quantitative evaluation. The choice of another emissivity (which was not measured 
here) would change only the systematic error and not the standard deviation, as 
141 
eu gus 
E > < 
v X a 0 SCO OR es —