ipment stages 
i dates since 
:al Stage 
’hase (70-80) 
i’hase (85-95) 
se(125-135) 
ding to 22nd 
been used to 
led out of 1: 
omenological 
INES. 
three of the 
/n in figure 2. 
DC). A clear 
r types of the 
iage rice and 
significant. 
: reflectivities 
classification. 
• of bordering 
titude in the 
in categories, 
an and citric - 
to the smaller 
nds chosen as 
other band 
ements. The 
iluation of the 
g prior to per 
h class values 
hnique of the 
it ion. But the 
VI instead of 
ited a better 
e out of the 
U ST 
160 180 200 
E X 
ndex images. 
7 th AUGUST 
Figure 4. Spectral response, in digital counts (DC), of several 
rice varieties in bands 4 (B4) and 5 (B5). 
procedure applied to the smoothed VI images of the three days 
considered. This technique gives good results for July and 
August. However, it is not applicable to the September image 
where the small differences in bands 4 and 3 between rice and 
citrics, figure 2, cause that the degradation of the image, VI, 
eliminates its possible application to the discrimination between 
rice and citric orchards. 
Test areas of 3 x 3 pixels are very well classified for each class 
with the smoothed VI image. However, smaller parcels, 2x2 
pixels, of urban or citric - vegetable zones are hardly 
distinguishable. Test area size of 2 x 2 pixels of urbanized class 
confuse themselves with those of the same ground size with 
short vegetated cover. 
The overall classification of the zone sets up 3730 ± 70 ha of 
rice, 1910 + 40 ha of citric - vegetables and 153 ± 4 ha of 
urbanized areas, 105 ± 2 ha are not classified. 
The image of 7th August for the area of La Albufera offers a 
greater distinction between the five varieties of rice studied.The 
image of September corresponding to the harvest phase is 
completely useless as it is expected from the study of the Sueca 
area. Finally, the image of July which corresponds to a less 
advanced period of the maturation phase is less significant. 
Figure 4 shows a measurement of reflectivity expressed in digital 
counts (DC) of the different varieties. 
The most significant measurements, as far as the distinction of 
varieties is concerned, are obtained in bands 4 and 5. Thus, 
figure 5 is a bidimensional diagram constructed from the image 
of August where the spectral classes, Bahia y Sequial are 
separated. It does not seem to be possible to separate the other 
varieties but its proportion is not significant, less than 1%. 
A first estimation of the zone of La Albufera gives overall 
results of culture of the variety Sequial higher than those 
supplied by the local public institutions. The results are partially 
obscured by the different degree of maturation of the varieties. 
We understand that a more complete treatment of the 
measurements (for example, of principal components [Richards, 
1984]) in an image corresponding to the maturation phase, could 
give a separation between the varieties Bahia and Sequial. We are 
trying now to locate test parcels (those of which inventary does 
not exist) to be able to evaluate the results of principal 
components analysis. 
4. REFERENCES 
Atkinson, P., Cushnie, J.L.,Townshend, J.R.G. & Wilson, A. 
1985. Improving Thematic Mapper land cover classification 
using filtered data. Int. J. Remote Sensing,Vol. 6, № 6, 955 
- 961. 
Berg, A. & Paretesi, S. G. de, 1980. Some significant results of 
a remote sensing experiment under European conditions 
(Agreste Project).Int. J. Remote Sensing, i, 3. 
Commission des Communautés Européenes (CCE). 
Etablissement d'Ispra, 1982 - 1984. Prévision des 
7th AUGUST 
B 
A 
N 
D 
4 
120 . 
118 • 
116 ■ 
114 • 
. o RUBINO 
112 • 
110 • 
. o UDO 
108 - 
106 ■ 
. o BAHIA 
104 ■ 
102 • 
100 • 
1 1 1— 
PIERINA o 
o 
SEQUIAL 
h 1 1 1 1 1 
54 55 56 57 58 59 60 61 62 63 64 
BAND 5 
Figure 5. Spectral response, in bands 4 (B4) and 5 (B5) of 
several rice varieties, in digital counts (DC). 
productions rizicolespar télédétection dans le bassin du Niger 
en amount du Delta Central Nigérien. Project F. E. D. 
№ 4507.03397.32. 
Commission des Communautés Européenes (CCE). 
Etablissement d'Ispra, 1978. Agreste Project: Agricultural 
resources investigations in northern Italy and southern France. 
Patel, W. K., Singh, T. P., Balder Saha & Patel, M. S.,1985. 
Spectral response of rice crop and its relation to yield and yield 
atributes. Int. J. Remote Sensing,Vol. 6, № 5, 657 - 664. 
Richards, J. A., 1984. Thematic mapping from multitemporal 
image data using the principal components transformation. 
Remote Sensing of Environment, 16: 35 - 46. 
Toll, D. L., 1985. Effect of Landsat Thematic Mapper sensor 
parameters on land cover classification. Remote Sensing of 
Environment 17: 129 - 140. 
ACKNOWLEDGMENTS 
We would like to thank EARTHNET - NPOC SPAIN for supplying 
the Landsat 5 (TM) tapes and the FEDERACION DE 
AGRICULTORES ARROCEROS DE ESPAÑA for their assessment.