The analysis of the resulting image of the ratio in TM4 (1993) 
and TM4(1984) led to the following relation : 
Objects with unchangeable reflectance have a value close to 1; 
Objects with significant change have a ratio different from 1. 
Two thresholds of change Tmin and Tmax Were determined on 
the histogram of the ratio image using mean digital number of 
some objects with significant change between the dates of 
acquisition. Hence, the area of change was extracted regarding 
the following relation : 
- values comprised between the DNmin and Tmin represent 
objects with decreasing reflectance. Values comprised between 
the Tmax and DNpax -represent objects with increasing 
reflectance. 
- non change area is represented by values comprised between 
Tmin and T max- 
3.2.1. Results 
According to the thresholds of change and the relation given 
above, a forest change image corresponding to TM4 band was 
obtained (see figure below). 
It is however necessary to find out a relationship between 
reflectance change and ground truth. 
Figure. change detection map 
175200 180900 384900 
| 
  
GR decrmaced vaties 
mcreased vabies 
272000 
C non-change area 
2689000 
264.000 
  
  
  
  
  
4. CONCLUSION 
The adopted model takes into account the topographic features 
of each pixel. However, slope and aspect maps were used. The 
estimation of the coefficients of the linear regression led to 
atmospherically comparable images. Then a ratio between the 
corrected images was calculated to find out significant change 
of each pixel. 
The change map obtained, therefore, allowed us to localise the 
change occurred in the study area. However, this was not 
sufficient enough to say that the change really happened unless 
it is confirmed by ground truth. This study constitute a topic of 
further research. 
REFERENCE 
Andrieu B., 1982, Evaluation des données fournies par 
télédétection aérospatiale pour l’étude multithématique d’un 
massif forestier, Thèse de Docteur-Ingénieur de l’Université de 
Pierre et Marie Curie, Paris VI, 189p. 
Beaubien J.G., 1990, Application de la télédétection spatiale 
à la cartographie des formations forestières et au suivi de leur 
évolution, IFN, Options Méditerranéennes Série A N°4, 1990, 
pp. 145-149. 
Cavayas F., 1984 Correction radiométrique des effets 
topographiques sur des images Landsat de territoires forestiers, 
Thèse de Ph.D de l’Ecole des Gradués, Université Laval. 177 
p. 
Chamignon C., 1990,Evaluation comparée des satellites à 
haute résolution spatiale en zone de moyenne et haute 
montagne méditerranéenne, Thèse de Doctorat de l’Université 
de Droit, d'Economie et des Sciences d' Aix-Marseille III, 377 
p. 
Danson F. M., Curran P. J., 1993,Factors Affecting the 
Remotly Sensed Response of Coniferous Forest Plantations, 
Remote Sensing Environment, N?43, 1993, pp. 55-65 
Durrieu S., 1994, Utilisation de la télédétection satellitaire 
pour la mise à jours de la carte des types de peuplements de 
l’IFN: Application à une région forestière diversifiée de 
moyenne montagne, Thèse de Doctorat de l’Ecole Nationale de 
Génie Rural des Eaux et des Forêts 202 p. 
Hakan O., 1994, Reflectance Calibration of Thematic Mapper 
data for Forest Change Detection, Swedish University of 
Agricultural Sciences (Part IV), pp. 1-21. 
Hakan O., 1994, Relative Calibrated Landsat TM data for 
Standwase Change Detection in Forestry, A pilot study on 
Scots pine infected by Gremmeniella abietina, Swedish 
University of Agricultural Sciences(Part-I), Remote Sensing 
Laboratory, S-901 83 Umeä, pp. 1-9. 
Kusaka T., Kakehi T., Ootuka M., 1996, Change detection 
using relative atmospheric correction of satellite images at 
different times, XVIII ISPRS-Congress, Vienna, 9-19 july 96, 
Article, 5 p. 
678 International Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998 
AB 
Alg 
oft 
Thi 
bas 
real 
ste] 
ma 
Fin 
GIS 
RE 
Les 
néc 
Cet 
util 
et c 
sur 
étu 
mé 
mé 
Il e 
(Ca 
l'aj 
MC