inbul 2004 
ti-temporal 
toring in 
ng, 22(16), 
fodulation: 
Improving 
e Sensing, 
aker, J. R. 
1. Remote 
Integration: 
grammetric 
Sensing & 
Sons, Inc, 
ogg, D. H. 
[he Use of 
ification of 
Geographic 
nentation of 
sification of 
& Remote 
COMPARING DIFFERENT SATELLITE IMAGE CLASSIFICATION METHODS: 
AN APPLICATION IN AYVALIK DISTRICT,WESTERN TURKEY. 
Aykut AKGÜN*"', A.Hüsnü ERONAT? and Necdet TÜRK"- (aykut.akgun(deu.edu.tr) 
" Dokuz Eylul University, Department of Geological Engineering, Building-A, IZMIR 
b > . > ; S . Ai. : er ; 
Dokuz Eylul University, Institute of Marine Sciences and Technologies, IZMIR 
KEYWORDS: Classification, GIS, Image, Satellite, Landsat 
ABSTRACT 
The different satellite image classification methods were compared using the satellite images of the Ayvalik district located 
- ^^ . . 2 . . m . . . . 
on the western coast of Turkey covering approximately 560 km”. For this purpose, landuse classification of the Investigation 
area was made by different supervised image classification procedures and the results were compared with one another. 
Landsat 7 ETM+ satellite image, IDRISI Klimanjaro image processing and the GIS package were used in this study. Of the 
classified images, the maximum likelihood method is found to be more applicable and reliable for the satellite image 
classification purposes. While the minimum distance method has given more reliable results than the linear discriminant 
procedures, the parellelpiped method is found to give the least reliable results compared to the other methods. 
1. INTRODUCTION 
Image classification is an important part of the remote sensing, 
image analysis and patern recognation. In some instances, the 
classification itself may be the object of the analysis. For 
example, classification of landuse from remotely sensed data 
produces a map like image as the final product of the analysis 
(Campbell 2002). The image classification therefore forms an 
important tool for examination of the digital images. 
The term classifier refers loosely to a computer program that 
implements a specific procedure for image classification 
(Campbell 2002). The analyst must select a classification 
method that will best accomplish a specific task. At present, it 
is not possible to state which classifier is best for all situation as 
the characteristic of each image and the circumstances for each 
study vary so greatly. Therefore, it is essential that each analyst 
understand the alternative strategies for image classification so 
that he or she may be prepared to select the most appropriate 
classifier for the task in hand. 
At present, there are different image classification procedures 
used for different purposes by various researchers (Butera 
1983, Ernst and Hoffer 1979, Lo and Watson 1998, 
Ozesmi&Bauer 2002, Dean&Smith 2003, Pal&Mather 2003, 
Liu et al 2002) . These techniques are distinguished in two main 
Ways as supervised and unsupervised classifications. 
Additionally, supervised classification has different sub- 
1091 
classification methods which are named as parellelpiped , 
maximum likelihood, minimum distances and Fisher classifier 
methods.These methods are named as Hard Classifier. 
In this study, the Ayvalik district located on the western coast 
of Turkey (Figure 1) was selected as a study area covering 
approximately 560 km? for comparing the 
  
  
  
^7 
- 
M il um AAN) "REDREMIT |. > 
/ ; / CT Balikesir w ay 
/ EDREMIT BUY ^" 
#3 N 
- 
> x 
™ 
Y 
\ 
> 
  
  
  
    
  
\ LA 
\ 25 N 
^ -— 1 
x Ar x. - 
^ N 4 e 
AEGEAN SEA - at M res June” 
RL mer BERGAMA 
4 
’ 
/ 
+4 / 
; í = Scale 
lumnirway-4—--7^ ^ Qt TORKM 
7" \ 
X 
  
  
  
  
  
  
Figure 1. Location map of the study area