2004 : : : ; : ; 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
  
  
d so 
the 
the 
“VA 
ered 
atus; 
ation 
1 to 
ry in 
local 
' the 
f the 
VA 
bove 
g of 
v hole 
This 
s the 
oe in 
rtage 
d-use 
1 the 
n in 
local 
area 
; that 
| belt 
other 
than 
  
:Satisfaction 
:Satisfaction in the whole local area 
:Shortage in the whole local area 
:Shortage in the current land use 
:Relatively wide dimension in the local area 
:Relatively narrow dimension in the local area 
Figure 4. Result of the decision tree method developed in this study. The result contains 6 categories concerned with status of 
vegetation cover. 
AIVA, it is necessary to consider it to be conserved and to 
become a factor composing the linear of VA. 
3. RESULTS AND DISCUSSION 
3.1 Results of Application of the Decision Tree Method 
Figure 4 shows the result of application of the decision tree 
method. It seems that there are large areas of "Satisfaction" 
and "Satisfaction in the whole local area" in the northwest 
and northeast part of this site. While the northwest part 
contains large residential areas (new town type), the 
northeast part has a lot of agricultural lands. Most of center 
areas have a tendency to belong to "Shortage in the whole 
local area” colored with white. However, there are many parts 
of "Relatively wide dimension in the local area” colored with 
cyan around the white areas. These cyan areas seem to be 
situated along linear structures: confluent, road, and railroad. 
It is suggested that because these linear structures include 
greenery; trees lining a street, green lands along riverbeds 
etc., they may play important role in greening in the local 
area. 
3.2 Application of the Results 
Through interpreting the result of the decision tree method, 
some suggestions about greening planning are derived. This 
section examines briefly the interpretation of the result. 
3.2.1 "Shortage in the current land use" + "Shortage in the 
whole local area": Both "Shortage in the current land use" 
colored with white and "Shortage in the whole local area" 
colored with red on Figure 4 mean the shortage of VA. At 
the place where the white area and the red area are adjacent 
to each other, however, there area possibility to increase VA 
in a unpaved land use area. This increasing may become 
feasible since it is not necessary to change the current status 
of land use. 
3.2.2 "Shortage in the current land use" + "Relatively wide 
dimension in the local area": "Relatively wide dimension in 
the local area" colored with cyan reveals the start point of 
greening to make the lineation of the VA. In case that the 
cyan area and the red area are adjacent to each other, it may 
assume a higher priority for achieving the standard of 
vegetation cover ratio. 
4. CONCLUSIONS 
In this study, the decision tree method concerning 
relationship between land use and land cover was developed. 
The paved/unpaved information as a land use data and the 
vegetation-covered/uncovered areas as a land cover data were 
applied to this method. It was implied that the results of this 
method might provide 6 suggestions for greenery planning. 
REFERENCES 
Hirano, Y., Yasuoka, Y., and Shibasaki, R., 2002. Pragmatic 
approach for estimation of vegetation cover ratio in urban 
area using NDVI. Journal of The Remote Sensing Society of 
Japan, 22 (2) , pp.163-174. Cin Japanese) 
Kumagai, K., Ishizawa, H., and Kawakatsu, Y., 2003. 
Analysis of the distribution of green tracts using satellite 
remotely sensed data. Proceedings of International Work 
"Shop on Monitoring and Modeling of Global Environmental 
Change, ISPRS WG VII/6, pp.H7.1-H7.4 
Kumagai, K. and Saito, G. 2001. Analysis of the 
urbanization using the extent and distribution of land-cover 
changes. Proceedings of the 22nd Asian Conference on 
Remote Sensing, pp.825-830. 
491