| 
ri 
1 
i 
i 
i 
N 
| 
1 
{i 
1 
d 
RR BEI la Re ST ata 
  
Tables 2 and 3 present the relative variance of the classified 
images (original and corrected) with various values for the 
confidence radius (R = 0.100, 0.125, 0.150, 0.175, 0.200). This 
is the principal input parameter of the classification method 
used. 
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
R Cl. Relative Variance 
TM2 T™3 | TM4 | TMS | General 
0.100 | 20 70.8 765) 57.3 91.1 64.7 
0.125 } 10 55.3 56.0 9.0 77.8 48.3 
0.150 9 54.8 55.1 9.0 75.7 47.3 
0.175 7 30.3 50.6 8.3 69.0 43.0 
0.200 6 45.6 47.8 8.1 69.0 42.5 
Table 2 - Original image 
R CI. Relative Variance 
TM2 T™3 | TM4 | TMS | General 
0.100 | 39 79.1 84.6 | 82.7 89.1 70.2 
0.125 | 21 74.3 78.4 | 67.2 85.2 64.0 
0.150 | 19 75.1 80.3 | 67.0 84.3 64.1 
0.175 | 14 70.7 76.1 59.0 77.9 58.6 
0.200| 11 64.6 64.3 9.8 75.1 44.4 
  
  
  
  
  
  
  
  
  
Table 3 - Corrected image 
We have always more classes in the corrected image than in 
the original image. The relative variance is always greater in 
the second case, mainly for the band TMA that is very 
important to the distinction of vegetal surfaces. The visual 
analysis confirms the better classification of the corrected 
image mainly for the vegetal surfaces that is our main interest. 
5. FINAL DISCUSSIONS 
This paper has done a general presentation of our work on 
atmospheric correction of satellite images in a tropical region. 
One of the principal characteristics that can be seen in this 
description is the work on various phases like the collecting of 
input data, the processing of the image with a software 
developed by us and the evaluation of the importance of the 
atmospheric correction in our situation. 
The influence of the atmospheric effects in two typical 
applications of the satellite images as showed in the section 4 
illustrates the importance of the work that has increased after 
the acquisition of a Noaa-AVHRR antenna on December 1994. 
The necessity to have good input data justifies the work on 
acquisition of atmospheric parameters. It is very important also 
to have a software developed according to our conditions of 
data, applications and processing, mainly considering the low 
disposability of such kind of systems around the world. 
It is important to emphasize that this work is inside a research 
program of the Campinas State University (UNICAMD) about 
the application of satellite images in the agriculture. The 
European Unit, EMBRAPA and CNPQ support this program. 
6. ACKNOWLEDGMENTS 
The author thanks all people from Campinas State University 
(mainly — Dr.P.C.Bezerra, Dr.H.S.Pinto, Msc.E.Hamada, 
Msc.G.Q.Pellegrino, MSc.C.A.S.Almeida) and INRA (mainly 
Dr.G.Guyot, Dr.B.Seguin and Dr.X.F.Gu) for all help received 
during these years of work. 
7. REFERENCES 
Carpenter, G.A., Grossberg, S., 1987. ART2: Self-organization 
of stable category recognition codes for analog input patterns. 
Applied Optics, N.26, pp.4919-4930. 
Holben, B.N., Eck, T.F., 1990. Precipitable water in the Sahel 
Measured using sun photometry. Agricultural and Forest 
Meteorology, N.52, pp.95-107. 
Lazutim, L.L., 1993. Personal communication. 
Lazutim, L.L., Bezerra, P.C., Fagnani, M.A., Pinto, H.S., 
Martin, L.M., Silva Mello, M.G., Turtelli Jr., A., Zhakov, V., 
Zullo Jr., J., 1994. Surface Ozone Study in Campinas Industrial 
Region of Brazil. Atmospheric Environment, England, 10p. 
London, J., Bojkov, R.J., Oltmans, S., Kelley, J.I., 1976. Atlas 
of the global distribution of total ozone July 1957-June 1967. 
NCAR Technical Note, N.113, USA. 
McClatchey, R.A., Fenn, R.W., Selby, J.E.A., Volz, F.E., 
Garing, J.S., 1971. Optical properties of the atmosphere. 
AFCRL-TR-71-0279, Environment Research Paper, N.354, 
L.G. Hancom Field, Bedford, Massachusetts. 
Pao, Y.H., 1989. Adaptive Pattern Recognition and Neural 
Networks. Addison-Wesley, 309p. 
Reagan, J.A., Pilewskie, P.A., Scott-Flemming, LC., Herman, 
B.M., Ben-David, A., 1987. Extrapolation of earth-based solar 
irradiance measurements to exoatmospheric levels for broad- 
band and selected absorption-band observations. IEEE 
Transactions on Geoscience and Remote Sensing, V.GE25, 
N.5, pp.647-653. 
Rouse, J.W., Haas, R.H., Schell, J.A., Deering, D.W., Harlan, 
J.C., 1974. Monitoring the vernal advancement and 
retrogradation of natural vegetation. Nasa/GSFC Type III, 
Final Report, Greenbelt, MD, 371p. 
Tanré, D., Deroo, C., Herman, M., Morcrette, J.J., Perbos, J., 
Deschamps, P.Y., 1990. Description of a computer code to 
simulate the satellite signal in the solar spectrum: The 5S code. 
International Journal of Remote Sensing, 11(4), pp.659-668. 
Tucker, C.J., 1979. Red and photographic infrared linear 
combinations for monitoring vegetation. Remote Sensing of 
Environment, N.8, pp.127-150. 
Zullo Jr., J., Gu, X.F., Guyot, G., 1994. Estimation de 
l'épaisseur optique des aérosols à partir de la mesure de 
l'éclairement spectral global. In: Colloque International de 
Signatures Spectrales d'Objets en Télédétection, Val d'Isère, 
17-21 January, pp.173-178. 
834 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996