stanbul 2004 International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 
——— 
  
Considering the different texture methods independently, it 
  
  
  
  
  
  
  
  
a cannot be stated that there is a universal method that is best for Class Producer's User's 
uem i all cases, since the results seem to depend on the type of accuracy accuracy 
rr a problem treated. However, they are usually better when Citrus orchards 80 07. 86.18 s 
| ; T he ; ; : 4 . 60.10 
statistical coocurrence features are used. The combination of New Urban 88.09 92 23 
e statistical variables with any of the other methods, energy - ve T 
these ; 3 ik 2 m IN ; eneras Horticulture 86.38 87.66 
: | filters, Gabor filters or wavelets, produce a significant increase > 
ERAN : Hy Tr A ^ i Old Urban 89.04 94.70 
trasontss in the overall accuracy levels, especially with the latter. This is Residential : 
o - side 3 
problably due to the complementary condition of the methods esae pa 9s 
; ; . s ] C 5 
based on filtering with respect to the direct statistical method Industriel 2.10 $3.99 
  
  
based on the GLCM. It is interesting to note that using only 
three. Gabor filters (three features) it is possible to obtain Table 1. Accuracy percentages of the classification of the urban 
and level of relatively good classification results. area using all the texture features (4 methods) combined. 
of variables: 
ginal image. 
; In the forest arcas, the texture classification provides accurate 
om details of 
results in those classes where there are mixed spectral 
responses, such as reforestation, and where the density of 
vegetation is a crucial factor, such as high, mid and low-density 
forest. Some examples are shown in figure 4. 
    
S M 
ess was the : ^ 
«ture samples E 
ting set, both ^i 
the different ai 
selection of à 
ariables were 
esults of the 
cy, are shown 
ES} Citrus Orchards Old Urban IRSE Residential 
[| Horticulture [E58 New Urban [7] Industrial 
Figure 5. Texture classification of a detail image of the urban 
f most of the area. 
to the only 
ctral bands of 
[enorca). The 
ips of texture 
3.4 Spectral vs. texture classification 
The classification of forest area 3 (Menorca) was done in two 
steps. In the first step, the non-vegetation classes (residential 
and sea) were masked out by taking advantage of the spectral 
and radiometric properties of the QuickBird multispectral 
image. The sea was masked by directly thresholding the 
infrared band, and the residential areas (also including roads 
and cliffs) were extracted by thresholding the third principal 
component of the four bands. This is easily achieved using 
these images, due to their high radiometric resolution (11 bits). 
Once the two masks had been applied over the panchromatic 
image, the second step consisted of the vegetation classification 
of the remaining areas. In addition, this comparative process of 
classification was carried out using both texture and spectral 
bands. Table 2 shows the comparative results in terms of 
producer's and user's accuracies. 
  
Figure 4. Detail images of texture classification of mixed areas 
with reforested and mid-density forest (above); and three 
different levels of forest density (below). 
Regarding the urban application, there are some classes that are 
accurately classified using texture methods, such as residential 
areas and old urban areas, but there are many commission errors 
(34%) in the industrial class. It is difficult to create a 
representative texture signature of this area, probably because 
the spatial resolution used is not aproppiate for this class. Table 
ST 1 | shows the specific accuracy levels for the different urban 
classes, and figure 5 a detail of the classified image. 
  
  
  
  
  
  
  
  
  
ST2 
ST > 
\N MULTISPECTRAL TEXTURES MS+TEXTURES 
r ME CLASS Producer's User's Producer's User's Producer's User's 
Accuracy Accuracy Accuracy Accuracy Accuracy Accuracy 
Dense forest 54.11 57.02 58.00 78.38 53.67 82.92 
Shrubs 62.26 55.39 88.46 94.76 88.90 94.2] 
Pasture-cereal 99.78 99.74 92.07 92.11 96.76 96.21 
Scattered trees 41.96 42.25 85.27 75.22 87.00 75.67 
Forest-shrub 21.45 25.56 73.71 54.79 76.10 52.47 
Weeds 61.80 58.56 90.36 89.70 94.78 95.65 
xture variables Fallow 98.69 97.42 87.34 91.02 97.13 100 
  
  
  
  
  
  
  
  
  
Table 2. Results of the classification of Menorca using spectral variables, texture variables and a combination of both. 
1113