bul 2004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B-YF. Istanbul 2004
of Figure 6.
ney k, Since the acquired image is produced from a spatially coherent
ncerned uniform radiation source, the interferogram extracted from each
horizontal line is, with good approximation, equal to the
interferogram of the energy from a fixed pixel. In a different
wording, it is not necessary to acquire a complete set of frames
(data-cube) to reconstruct the full interferogram of the
(6) concerned pixel.
In Figure 7 and Figure 8 the average interferograms have been
plotted.
we can 1800 :
— raw inteferogram
sing is 1600
intity is Hae
1200 |
000 - |
7 800
HU WI { 1!
eoo HI | |
[BY the |
; 400
olution,
ratory . . : ; ; SE 200
: Figure 5. Raw image (in gray scale) obtained illuminating a 5
| He-Ne double planar diffuser with a 600W halogen lamp.
)btain a 0 200 400 . 600 800 1000
pixel
field at -As can be seen, the number of interference fringes is smaller
than in the previous image due to the broad spectral content of
omplete the radiation source. Moreover, the strong central fringe which
ing the corresponds to null interference is black, i.e. for the rays coming ; : S:
T p S : t aem te y 8 Figure 7. Interferogram averaged over all the columns of image
along the instrument optical axis their phase-delay is . of Fioure 4
ource 1s In Figure 6 the same lamp has been seen trough an interferential "
filter with bandwidth of 10 nm centered at 640 nm. EM
)
3 54 — raw interferogram
2 500 >
x
LE;
" : 400
€21
8 3 300 Wii,
d 3
3
5
i 200
T
up 100
E
TJ
318 0
ig 0 200 400 iv 600 800 1000
al pixel
ME
14
Hi
+
$ Figure 8. Interferogram average over all the columns of image
€
4
r
Se
3
OS m too
| i The pre-processed interferogram should have a null-mean,
Rr ah starting and ending tails approaching to zero, and any optical
artefact removed. In order to achieve these characteristics, we
ting a "igure image (i scale ined illuminating a :
Ade is Fi S. Ras Image fn EE pa have pre-processed our data according to the following scheme
age is double planar diffuser with a 600W halogen lamp seen through (Bard cel and Pip, 2001).
es ol an interferential filter centered at 640 nm and wide 10 nm. 2 :
1) dark signal subtraction (to account for bias and noise in
; : di > detector electronic stage):
Reducing the spectral content of the radiation source by tic deicetor elec ge);
's which filtering, the interferogram becomes more similar to that 2) instrument spatial response compensation (to remove
of these produced by a “quasi-monochromatic” signal. It is also evident plie pixels, hot and cold pixels, and fixed-pattern
> of the the effect of vignetting on the acquired frame. noise; | £60 monu ;
T is, in 3) geometrical and radiometric distortion correction (to
à of the 5. RETRIEVED SPECTRA remove effects of vignetting and spatial shift of the
fringes);
ith the From the frames depicted in Figure 4 and Figure 6, it has been 4) DC-offset subtraction to estimate the band-averaged
"igure 5. extracted the corresponding interferogram averaged over all energy reaching the detector as expressed by the term
image pixels. 19(k)/2 in (1);
133
