Vol. 
GIS 
ion. 
Jues 
74, 
Darvishzadeh Varchehi, Roshanak 
  
Change Detection for Urban Spatial Databases Using Remote Sensing and GIS 
Roshanak DARVISHZADEH 
Tehran municipality- Tehran GIS Centre (TGIC) 
P. O. Box 19575/597, No. 40, Badiee intersection, Aqdasieh Ave, Tehran, Iran 
darvish@itc.nl 
KEY WORDS: Change Detection, GIS, Remote Sensing, Urban. 
ABSTRACT 
The aim of this study was to establish a preferred methodology for updating of the spatial database and monitoring the 
changes of a settlement in Tanzania. In summary, change detection problems can be viewed as the consequence of an 
increased desire for up-to-date information about the development and changes in a GIS environment to guide urban 
planners, urban decision-makers, and managers. Two methods that have been used for change detection are: semi- 
automatic change detection using on-screen digitizing and digital classification and segmentation methods. 
On-screen digitising of changes shows that (as a pilot study covered only one aerial photograph) this method is fast, 
convenient, and (depend to the application) complete. 
The result of segmentation and classification for finding the changes in the settlements Dar Es Salaam, shows 7396 
accuracy when the result of on-screen digitizing is used as a reference map for evaluation and by using existing 
knowledge such as roads and object geometry contained in the GIS database. Classification and segmentation can be 
used to derive information from an aerial photograph. The nature, variety and usefulness of such an information would 
have to be considered according to the project at hand, i.e., the structure of the area, the type of available aerial 
photograph, and the type of information to be extracted. 
1 INTRODUCTION 
In Developing countries new expansion due to development of unplanned area and urbanisation causes maps to become 
out-dated rapidly. Therefore, there is a need to revise old maps using a fast, efficient and effective method. 
The problem addressed by this study is the lack of up-to-date and correct information about the development and 
changes of unplanned areas in Dar Es Salaam to guide urban planners, urban decision-makers and managers. Estimation 
by the City Council of Dar Es Salaam (UNDP, 1992), indicates that approximately 70 percent of the population live in 
unplanned settlements. These settlements are vast in extent and their population grows so fast that traditional data 
collection, processing and analysis techniques cannot provide planners and decision-makers with a correct view of 
them. The time interval between the production of the first map with the last map of a series may be too long. As a 
result a lot of new changes may have occurred again. Traditional map production processes are inadequate for this 
situation. 
To revise an old map, it is necessary to detect the changed object in the area. An old map, and old and new aerial 
photographs are important basic materials for change detection. There have been a number of processes that use 
analogue aerial photographs for change detection. An example is plotting the changes with a mechanical stereo-plotting 
machine. But these processes are time-consuming, laborious, and, consequently not economical. Finding a more 
appropriate solution to this problem has become a genuine challenge. However, because of the possibility that now 
exists to scan analogue aerial photographs, it is possible to utilise them in a more efficient way. This study has shown 
how digital aerial photographs can be used to maintain the spatial database and update maps and the relevant 
information for urban planning in a faster, cheaper, and more convenient way. 
2 DESCRIPTION OF THE METHOD 
Before extraction of new data from the new aerial photograph (1994), it is worthwhile to compare the existing 1992 
aerial photograph with the digital spatial database which has been produced from that photo. The reason for doing this 
is that the 1992 digital data will be used as reference data and the quality and accuracy of the source data should be 
ascertained before using it for further processing. This was done by overlaying the aerial photograph on the digital data 
and comparing some of their details using visual interpretation. (Darvishzadeh, 1997) 
  
International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000. 313