ata sets 
ed with a classi- 
comparison of 
listed there 
e corresponding 
et the classi- 
n May, cereals 
ses of cereals 
ntly better 
ata enhance- 
ificant extent. 
s not inconsi- 
w of certainty 
daries. 
al fields in 
lata with better 
Jleberlagerung von 
ing. in: Bild- 
Flector. in: 
, Los Angeles, 
APPLICATION OF COMPUTER AIDED ANALYSIS OF LANDSAT DATA TO 
LAND USE STUDIES IN BELGIUM 
The Ciergnon test site (Famenne) 
  
A.Ceusters,R.Gombeer, H.Gulinck,N.Sougnez,J.D'Hoore 
Laboratorium voor Bodemgenese en Bodemgeografie K.U.Leuven 
Centrum voor Landbouwkundige en Ekologische Interpretatie 
van Teledetektiegegevens (C.L.E.I.T. - I.W.O.N.L.) 
Summary 
The almost rectangular test site of 7 x 8 km (5.670 ha) is located in the 
Famenne, a small natural region between the Condroz and the Ardennes. 
Four land use categories have been retained: coniferous forest (Picea 
abies dominant), deciduous forest (Quercus, Carpinus), grassland and 
cultivated land. The two CCT's studied were recorded on 22/3/1973 and 
17/5/1976. 
Homogeneous test arcas the land use of which was correctly known were 
located by means of the NMAPW (ORSER) programme. and spectral signatures 
were statistically analysed by means of the STATS programme. This has 
shown that in the Famenne, both in March and in May band 5 is the most 
selective. 
Radiance ranges for each category in band 5 were then introduced in the 
EXIO programme, which produced a map according to preset radiance limits 
and a distribution statistics, which was in close agreement with data 
obtained by planimetry on the 1/25.000 NGI map. 
The supervised classification programme CANAL, which makes use of the 
four spectral bands, gave results that were rather similar to those pro- 
duced by EXIO. 
Introduction 
The application of computer aided analysis of remote sensing data to rural 
economy and regional planning problems presents some obvious advantages. 
Besides the fact that much of the work can be automatized,maplike repre- 
sentations can be generated which allow a rather exact positioning of pi- 
xels on the terrain, at least in a well mapped country. ,Secondly,programmes 
can be aimed at specific land use features at propitious times of the year, 
and then yield good approximations of spatial distribution at the time of 
the overflight.