Orientations between 35-55 and 125-145 degrees from 
the flight direction are regarded as critical. 
For example, this explains the relatively poor pro 
portion of pure pixels in the site of St.Jan (SJ) 
although the fields are generally large. The domi 
nant orientation of the field boundaries is evident 
for the very elongated fields, but is also clear 
for the small, compact fields. In general, most 
fields in sites analysed in Flanders, show critical 
orientations. 
5. CONCLUSIONS 
Thematic image interpretation of areas such as occur 
in northern Belgium, will remain difficult even with 
the 2nd generation remote sensing satellites. The 
main reasons reside in the great diversity of land 
scape having a very fine and complex structure and 
varying over short distances in space. Important 
structural parameters proved to be the density of 
the landscape elements, the size and the compactness 
of the land use fields, the orientation of the field 
boundaries and the density of the biotic screens 
(especially the shadow giving tree- and hedgerows). 
All these determine the proportions of the pure and 
mixed pixels in the image for a given land use cate 
gory. Generally the proportion of the pure pixels 
remains very low and varies a lot between the dif 
ferent landscape types in the case studies in 
Flanders. 
Landclassification techniques using structural 
indicators permit a fast landscape typology. 
We suggest to use this for a stratified sampling 
of training areas. Further research will focus upon 
the imp!imentation ÿsuch a structural landscape 
typology in image classification and the elaboration 
of interpretation keys. 
REFERENCES 
Ackerson V. & Fish B. 1980. An evaluation of Land 
scape Units. Phot.Engin, and Rem.Sens.,vol.46,3, 
347-358. 
Antrop M. & Daels L. 1977. The extraction of soil 
information out of Remote Sensing documents. 
Gent, Pédologie, XXVIII,2,123-190, 10 tab., 21 fig. 
Antrop M. & Snacken F. 1981. Structure and dynamics 
of landscape systems. In : J.Drdos : Landscape 
Synthesis : Geoecological Foundations of the 
Complex Landscape Management. Bratislava, IGU- 
Working Group on Landscape Synthesis, Slovak Acad. 
Sciences, 10-30, 7 fig., 1 tab. 
Antrop M. 1982. The "Natural" way of visual image 
interpretation for landclassification and land 
scape planning. Toulouse, Actes du Symp.Intern, 
de la Comm.VII de la Soc.Intern, de Photogramm. 
et de Télédétection, ed.GDTA, Intern.Archives of 
ISPRS, Vol.24-VII/1,897-906, 1 tab., 2 fig. 
Antrop M. 1985. Télédétection et analyse du paysage. 
In : V.Berdoulay & M.Phipps (ed.), Paysage et 
système Ottawa, Editions de l'Université d'Ottawa. 
Antrop M. 1982. Group Truth Collection for a Visual 
Interpretation of SAR-580 imagery on the B1 Site 
in Belgium. New York, IEEE Intern.Geoscience & 
Remote Sensing Symp. (Munich), Digest, Vol.I.WA- 
5,5.1-5.6. 
Antrop M. 1983. Inventoring and monitoring of land 
scape as a natural and cultural resource. 
Paris, ESA, sp-188, Proc.EARSel/ESA Symp.on 
Remote Sensing Applications for Environment 
Studies, 26-28 april 1982, Brussel, 105-113. 
Antrop M. 1983. Structural analysis of rural land 
scapes. Liege, IGU-Comm.Rural Development, Collo 
que Intern.d'Amenagement Rural, 4-11 sept. 1983, 
Vol.II, 1-31 
Antrop M. 1985. Analyse texturale du paysage par 
interprétation visuelle d'images Landsat (MSS7) et 
de photographies aériènnes. Hommes et Terres du 
Nord,3, 162-168. 
Aurelio V., et al. 1984. Land Use Evaluation by SPOT 
imagery on the Basilicata Test Site. In : P.Rei- 
niger.ed., SPOT Simulation Europe 82, 15-42. 
Curran P. 1985. Principles of remote sensing. 
London, Longman, 282 pp. 
Flouzat G. 1982. Modélisation de la compréhension 
visuelle des images fr télédétection : essai de 
simulation numérique de la photointerprétation 
analytique. Toulouse, Actes du Simp.Intern.de la 
Comm.VII de la Soc.Intern, de Photogr. et Teled., 
vol.2, 7-24 pp. 
Girard C.M. & GIRARD M.C. 1973. Interprétation des 
photographies-ballon : comparaison avec les peti 
tes régions agricoles. Bull.Soc.Française de 
Photogramm., 23-36. 
Girard C.M. & GIRARD M.C. 1975. Comparaison de 
photographies aériènnes et des images à haute 
altitude. Paris, rapport Inst.Nat.Agronomique 
Paris-Grignon, 5. 
Girard C.M. & Girard M.C. 1985. Interprétation du 
paysage à petite échèlle à partir de clichés de 
la chambre métrique Spacelab. approche botanique 
et pédologique.Bul1.S.F.P.T., nr.99-3, 41-51. 
Henderson F. 1980. Effects of interpretation techni 
ques on land-use mapping accuracy. Photogramm. 
Engin, and Remote Sensing, vol.46, nr.3, 359-367. 
Howard J. & Mitchell C. 1980 - Phyto-geomorphic 
classification of the landscape. Geoforum, vol. 2, 
35-106. 
Kilchenmann A. 1973. Character Analysis for Nominal 
Values. A Method for the Analysis of Qualitative 
Geographical Attributes Based on an Information 
Theory Model. Geoforum, 15, 33-45. 
Lam Tran Chau 1985. Some sampling methods for esti 
mating areas on maps. I.T.C. Journal,2, 113-120. 
Lillesand T. & Kiefer R. 1979. Remote Sensing and 
Image Interpretation. New York, Wiley. 
Mabbutt J. 1968. Review of Concepts of Land Clas 
sification. In : G.Stewart (ed.) Land evaluation, 
Melbourne, Me.Millan of Australia, 11-28. 
Reiniger P. (ed.) 1984. SPOT Simulation Europe 82 - 
Final Report. Ispra, JRC-Comm.E.C., SA/I.04.E2.84. 
11, 258. 
Townshend J. (ed.) 1981. Terrain Analysis and Remote 
Sensing. London, Allen & Unwin. 
Vandecasteele I. 1979. Landschapsclassificatie toege- 
past op België d.m.v. textuuranalyse van Landsat- 
beelden (MSS7). Gent, RUG, lie.thesis, 102. pp. 
Vander Haegen H. 1982. Het bodemgebruik in België 
en de evolutie ervan sinds 1834 volgens de kada- 
strale gegevens. Brussel, N.I.S., Statistische 
Studien, nr.65, 33pp. 
Van Genderen J. Lock B. Vass P. 1978. Remote Sensing: 
Statistical Testing of Thematic Map Accuracy. 
Remote Sensing of Environment, 7, 3-14 
Vauzelle M. 1982. La télédétection pour 1'approche 
des unités paysagiques en milieu bocager. Applica 
tions aux Monts d'Ambazac (Limousin-France). 
Bull.S.F.P.T., nr.88, 7-18. 
Webster R. & Beckett P. 1970. Terrain classification 
and evaluation using air-photography : a review 
of recent work at Oxford. Photogramm.Engin., 
vol.26, 51-75 
Symposiui 
Interprc 
Helmut Bei 
Department ft 
Manfred F. 
Institute for It 
ABSTRACT: Bas 
of combining 
use a computi 
explanations 
RESUME: A la 
discutés des 
des cartes gé 
pour reclassi 
des resultant 
1 INTRODUCT 
In correspo 
in this 
comprehensi 
remote se 
comparing 
geoscientif 
sensing in 
methods. 
to point ou 
interpretat 
their comb: 
data as we 
taken as a i 
a specific 
rectly esl 
geoscientif: 
area close 1 
Austria's 1 
tree-line (! 
made to r« 
thematic Is 
Figure 1. Lo 
8