International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004 Ini
nter!
Further work in this subtle-change-detection with the aid of
parameters such as NDVI from hyperspectral sensors, colour pium
information from RGB, NIR sensors (ADS40), road network the T
map along with the geometries of roof. The obtained GIS e
information was used for an application in the area of property nde
tax office.
Chu :
Maps
on I
1241:
Figure 4. NDVI and thermal data fusion result. Lower NDVI
values correspond to higher temperature regions (man-made Bruz:
features and moving objects). Asura
remo
Though the remote sensing data covers different areas, for the Geos
purpose of describing the capability of sensors, we integrated
thermal and hyperspectral images (AISA) as well as thermal Fegar
(TABI) and NDVI obtained from AISA for Tokyo region. fusio
Where as to show the advantage of using multidate DSM classi
exclusively extracted from high-resolution ADS40 Aust
stereoimages for subtle change-detection study, data of Adel
Yokohama was utilized. A GIS analysis for the n D
change-detected areas was carried out manually as well as | EE DUM Fung
automatically by remote sensing and GIS data fusion for fixed b) Colour information Date: 2003 enm
i res p : ; Remc
property tax division of Yokohama city. The accuracy of the
results for various building nature is summarized in Table 2. Hall
data |
A. Situation B. Extracted | C. Extracted Ratio p
Manually Automatically % Heo
(against A) Noni
New 151 151 100 Sense
construction | Sensi
Re-built 32 31 97
Non-existence 101 100 99 [tthi,
Expansion 35 22 5] differ
Under 47 46 98 http:/
construction m.
Unknown 98 76 78
Total 465 417 90 Krish
2000.
Table 2. Information from remote sensing and GIS fusion LARA : for I:
(d) Colour information Date: 2003 Sensi,
5.3 Colour information for Building roof identification
Figure 5. Importance of Colour information of Roof Le 1
Figure 5 shows the importance of roof colour information in a Segm
high-resolution image. From the DSM roof geometries such as Our automating the processing of images for importing to GIS Trans
shape, size and orientation can be obtained. But a close has the advantageous such as that it is possible to enter 605-6
observation of roof features in figure 5. (a) to (c) revealed that remotely sensed data into GIS in a routine economical fashion,
the geometry of the roof structures is similar but the spectral and then frequent updating of GIS data by fused remote sensing Mich:
signature (colour) is different. Therefore, the roof geometry imagery would improve the accuracy of applications. high-1
Automatic image processing of data fusion will provide a rapid vectoi
information could not be used in certain case. For the
change-detection purpose existing building data will be used to and application specific reconnaissance of high-resolution
find the difference along with an analysis of the colour images images from visible, IR, and thermal. A degree of automation Phol,
will also reveal the difference. through high-resolution and new remote sensing data and GIS in re
integration will facilitate the demand of government and public Interr
6. SUMMARY AND FUTURE WORK agency adoption of DSM, ADS40, TABI and AISA data. R
och:
* In this study, first an attempt was made to use DSM derived Our study presented a method of automating the processing of GIS
from pixel-pixel matching of ADS40 images for broad high-resolution imagery into a suitable form for import into use/cc
change-detection. Secondly, to develop a novel technology to GIS. In addition, the information (such as road and building) Over I
integrate multi-date, multi-resolution, and multi-sensor data for contained within the GIS can be used. The results obtained
the purpose of subtle change-detection analysis. from NDVT difference, Road data, roof geometry and colour Saraf,
image analysis will be combined with the DSM-difference da appro.
d. We also Intern
and final change-detection map could be achieve
904
