29 
lg the approaches to 
iication in GIPS/ALP. 
dcentre is identified 
fieldcentre pixel is 
. By comparing this 
ith that originally 
he change, if any, 
etected. The acreage 
:udy area is derived 
ata instead of the 
y at the GIS accuracy 
The results of the previous three parts are 
the inputs to the particular classifier to classify 
the cover type of each unknown pixel. And, the 
subsequent results of the newly derived land 
cover data can be used to update the archival 
land cover file in the GIS. 
4. UNIQUE ADVANTAGES OF GIPS/ALP 
This approach can dramatically reduce the data 
volume which has to be handled when spatial 
resolution is increased because the number of 
pixels of each crop field to be processed can 
be kept constant and the total number of fields 
for a defined size of area will remain unchanged 
in a certain period of time. The updating of 
polygon data file is not required so often as 
that of the content inside the polygon area. 
The classification accuracy will be considerably 
improved due to the purity of the pixel chosen 
and the high accessibility of ancillary data 
available in GIS. The field check strategy modified 
by this system can contribute the general accuracy, 
too. 
The geometric accuracy will be as good as that 
in the GIS which is claimed to be considerably 
higher than that can be derived from the most 
sophisticately corrected images because all the 
statistics related to geometry are estimated 
on basis of the data in geographic information 
system. 
5. CONCLUSION 
Identified Change 
New Cover Condition 
Cl N (no ch) 
C2 C (chngd) 
id truth verification 
be directed to those 
.xels or are detected 
.on can be accurately 
ated map, on which 
larked automatically 
include (i) statistic 
s, and (iii) images, 
1 both the geographic 
ensing system. 
assification 
strati the specific 
GIPS/ALP. A brief 
eographic information 
system are the two 
iLP. The geographic 
handling the fully 
note sensing system, 
e spatial data. 
tasks are involved, 
.ds with known land 
>es were taken, (ii) 
¡ponding the fields 
p, (iii) extracting 
modifying the image 
aining statistics 
ification algorithm, 
istribution of each 
n be estimated from 
GIS. 
which a value stands 
re derived from RSS 
ucture. 
Both geographic information 
sensing system are effective 
spatial data. GIPS/ALP takes 
special features to resolve 
image processing which have 
They are the data volume, classification accuracy 
and geometric accuracy. 
system and remote 
tools for handling 
advantage of their 
the bottlenecks in 
existed for years. 
REFERENCES 
Allan, J.A. 1984. The role and future of remote 
sensing. In proceedings of the 10th Anniversary 
International Conference on Remote Sensing. 
Remote Sensing Society, UK. p.23 30. 
Anderson, J.R., E.E. Hardy, J.T. Roach & R.E. 
Witmer 1976. A land use and land cover 
classification system for use with remote sensor 
data. Geological Survey Professional Paper 
964, USGS. 28p. 
Marble, D.F. & D.J. Peuquet 1983. Geographic 
information systems and remote sensing. In 
R.N. Cowell (ed.) Manual of remote sensing, 
vol.I, p.923 958. 
Tseng, W.T. & Q.C. Sung 1978. The evaluation 
of computer generated land use/land cover maps 
and the study of spectral characteristics of 
land covers. MRSO repot 088. Mining Research 
& Service Organization, ITRI. 39p.