Pozo, 
fold is 
Jp» FE- 
of the 
rately. 
inifold 
In par- 
he im- 
1struc- 
image 
sity of 
eed to 
ity ex- 
S way, 
e con- 
image 
ng the 
s done 
nenta- 
g. For 
| com- 
nages. 
rocess 
ortant 
which 
of the 
of the 
struc- 
| gra- 
. The 
retain 
ppen, 
' sub- 
ound- 
rithm 
othe- 
of the 
n de- 
econ- 
ction 
used 
is the 
othed 
yrma- 
od at 
thus, 
| het- 
cally 
s like 
omo- 
rs of 
ssing 
able. 
il 2004 
    
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004 
  
    
Figure 4: Details of the reconstructed images and their 
graylevel values' distribution. From top to bottom: de- 
tails of the original Spot image, details of the FRI and the 
RMI, respectively reconstructed with the fields v4; and vo. 
The fidelity to the original image can be adjusted, by grad- 
ually incorporating as many details of the original image as 
desired, so that the degree of smoothing can be controlled. 
For a deeper validation of the method, we should perform 
the algorithm on very high resolution images. 
REFERENCES 
Daubechies, I., 1992. Ten lectures on wavelets. CBMS- 
NSF Series in Appl. Math., Capital City Press, Philadel- 
phia (USA). 
Frisch, U., 1995. Turbulence. The Legacy of Kolmogorov. 
Cambridge University Press, Cambridge MA. 
Grazzini, J., Turiel, A. and Yahia, H., 2002. Entropy esti- 
mation and multiscale processing in meteorological satel- 
lite images. In: Proc. of ICPR, Vol. 3, pp. 764-768. 
Laporterie-Djean, F., Lopez-Ornelas, E. and Flouzat, G., 
2003. Pre-segmentation of high-resolution images thanks 
to the morphological pyramid. In: Proc. of IGARSS, 
Vol. 3, pp. 2042-2044. 
Marr, D., 1982. Vision. W.H. Freeman and Co., New York. 
      
    
1129 
  
     
  
Le ra 1 RE iur 
SR x eut nA : 1 
Use à 
EE NUS WO De 
— 
SF 
Figure 5: FRI obtained considering different MSM with 
varying density in the m From top to bottom: 13.9% 
and PSN R= 23.50dB, 22.3% md PSNR = 21.9345, 
32.5% and PSNR = 25.30dB (the PSNR were com- 
puted for the whole images). Note that the middle image 
is the same as the middle one displayed in Fig. 4 
Mather, P., 1995. Computer Processing of Remotely- 
Sensed Images: An Introduction. John Wiley & Sons, 
Chichester (UK). ond eq. 
Schiewe, J., 2002. Segmentation of high-resolution re- 
motely sensed data - concepts, applications and problems. 
Int. Arch. of Photogrammetry and Remote Sensing 34(4), 
pp. 380—385. 
Schowengerdt, R., 1997. Remote Sensing. Models and 
Methods for Image Processing. Academic Press, San 
Diego (USA). 
Turiel, A. and del Pozo, A., 2002. Reconstructing images 
from their most singular fractal set. IEEE Trans. on Im. 
Proc. 11, pp. 345-350. 
Turiel, A. and Parga, N., 2000. The multi-fractal structure 
of contrast changes in natural images: from sharp edges to 
textures. Neural Comp. 12, pp. 763—793. 
Yocki, D., 1995. Image merging and data fusion by means 
of the discrete 2D wavelet transform. J. Opt. Soc. Am. A 
12(9), pp. 1834-1841.