. Istanbul 2004 
1 designed and 
im for network 
network 
atures with the 
e arranged in a 
? identifies the 
mage has been 
compared to 
method that is 
ods based on 
ile is based on 
ular class. The 
s are equal for 
i| distributions. 
1sing MLC. 
cation method, 
| classification 
reference map 
. The obtained 
methods are 
rom a global 
ployed in this 
yrovement in 
method. 
cation method 
lihood method 
that the neural 
C method. As 
in the neural 
ent. Also, the 
ning sites than 
nprovement to 
ssification, the 
s compared to 
> MLP method 
is not trained 
MLC method 
fied all pixels 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
oss e F zia a 
20 40 60 80 100 120 140 160 
(c) Classified image using 
MLP 
  
(d) Classified Image using 
MLC 
Figure 4: The classification process, (a) The original image of IRS-1D satellite, (b) Segmented image using 
SOM method, (c) Classified image using MLP method, and (d) Classified image using MLC method. 
Using segmemation in this study, efficiency of the 
classification process has also been improved. This is due to 
the segmemation made before feature extraction to avoide 
time consuming process of redundant data processing in 
feature extraction stage. Also, with regards of efficiency of 
the process, Resilient back propagation that is generally much 
faster than the standard steepest descent algorithm has been 
applied. It also has the appropriate property that requires only 
a modest increase in memory requirements. This enables us 
to store the updated values for each weight and bias which is 
equivalent to storage of the gradient. 
7. REFRENCES 
Augusteijn. M.F., L.E. Clemens, and K.A. Shaw, 1995, 
"Performance evaluation of texture measures for ground 
cover identification in satellite images by means of a neural 
network classifier", /EEE Transactions on Geoscience and 
Remote Sensing, Vol. 33, No. 3, pp. 616-626. 
121 
Coifman, R.R.; M.V. Wickerhauser, 1992, "Entropy-based 
algorithms for best basis selection", IEEE Trans. on Inf. 
Theory, Vol. 38, 2, pp. 713-718. 
Godfery, K.R.L. and Y. Attikiouzel, 1992, Applying neural 
network to color image data compression, /EEE region 10 
conference, Tencon 92, Melbourne, Australia. 
Gonzales, R.C. and R.E. Woods, 1993. Digital Image 
Processing, Addison-Wesley. 
Kohonen, T., 1989, Self-organization and associative 
memory, Springer-Verlag, Berlin. 
Moreira, J., L.D.A. Fontoura Costa. "Neural-based color 
image segmentation and classification using self-organizing 
maps". 
Ohanian, P.P. and R.C. Dubes, 1992, "Performance 
Evaluation for Four Classes of Textural Features", Pattern 
recognition, 25(8), pp. 819-833.