lar(basic) 
ng element 
irection 
ng element 
ion 
element 2 
  
(x) 
1. In the case 
ont sizes, the 
ed according 
ments. WNED 
1 comparison 
| WNED can 
spurious edge 
element sizes, 
(js comparing 
tion. 
sctive in case 
one functions 
it can control 
d algorithm is 
| because it 
regardless of 
that it is able 
s aflood. 
nsional edge 
(mun. Image 
iig X, 1987. 
morphology. 
chine Intell, 
Multispectral 
ya using a 
  
cooperalive learning neural network. Proc. of 
IGARSS '94, Vol.2, pp.508-512. 
Kawamura M. and TSUJIKO Y., 1995. An approach 
to geographic pattern recognition using 
mathematical morphology. Proc. of IGARSS '95. 
lee J. 5% Horoick R and Shapiro LZ" 1987. 
Morphological edge detection. IEEE J. Robotics 
Automation, Vol.RA-3. 
Morages "P. i»*gnd. P9eScnaferc Roy ^v19879. 
Morphological filters-Part |. IEEE Trans. Acoust. 
Speech "Signal "Processing," Vol.ASSP-85, "No.8, 
1153-1169. 
  
Maragos P. 
Morphological 
Speech 
1170-1185. 
Matheron G., 
Signal 
and Schafer 
filters-Part Il. 
Processing, 
1975. Random 
Geometry. New York, Wiley. 
Figure 5 Results 
889 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B3. Vienna 1996 
R W.. 
FFE Trans. 
Vol.ASSP-35, 
Sets and 
19870. 
Acoust; 
No.8, 
Integral