International Archives of Photogrammetry and Remote Sensing. Vol. XXXII Part 7C2, UNISPACE El, Vienna, 1999 
52 
I5PR5 
UNISPACE III - ISPRS Workshop on 
“Resource Mapping from Space” 
9:00 am -12:00 pm, 22 July 1999, VIC Room B 
Vienna, Austria 
/-?m\ 
ISPRS 
SATELLITE REMOTE SENSING APPLICATION IN AGRICULTURE- 
CROP MONITORING, YIELD FORESCASTING AND ESTIMATION 
Cs. Ferencz, J. Lichtenberger, D. Hamar and P. Bognâr 
Space Research Group, Department of Geophysics, Eôtvôs University, Budapest, Hungaiy 
H-l 117 Budapest, Pâzmâny Péter sétâny 1/A. (H-1518 Budapest, Pf.32., Hungary). 
Tel: +36-1-372-2906, fax: +36-1-372-2927, 
E-mail: spacerg@sas.elte.hu 
ABSTRACT 
Crop monitoring, yield estimation and forecast are widely considered as a matter of strategic importance. This is particularly true for 
developing countries, where a cheap, effective and easily adaptable yield forecasting/estimating procedure for county/country level essential. 
A method based on satellite data that fulfill this criteria is presented beside a high resolution method that can supply regional/field data and 
the calibration of the basic method. The basic, i.e. robust method is based on low resolution (NOAA AVHRR) data. The benefit of using 
AVHRR data is its daily availability and low/no cost. The method of calibration use is based on Landsat TM data too. The procedures 
consists of the following main steps: 
□ preprocessing of raw satellite data: geometric and atmospheric correction, 
□ development of calibration functions (i.e. crop filters) only' one times for a given crop, 
□ calculation of special agricultural RS indeces for the given application, stress or disaster detection, 
□ forecasting and final estimation of crop yields. 
The results of tlie application of this method confirms of tire reliability, operational applicability and cheapness of tins procedures, which are 
important for the developing countries. 
1. INTRODUCTION 
As it is well known the main application fields of satellite remote 
sensing (RS) □ e.g. Avenasov, 1996 □ are the followings: 
meteorology; agriculture, investigation of canopy and soil; water, 
It is clear that the complex agricultural applications of satellite 
remote sensing (RS) use not only the agricultural canopy and soil 
investigations □ e.g. local, regional and global monitoring of 
canopy, soil, soil moisture and soil degradation □, however, they 
use other RS applications too, as environmental monitoring, 
meteorology and global changes, canopy-, crop- and land use 
mapping. This new data set could be an integrated part of GIS and 
can change the future agriculture, improve the stability of supply 
and trade. 
In the present □ and in the near future too □ an important 
parameter of applications is the necessary on-surface resolution of 
the used satellite data. The low resolution lias some inherent 
limitations parallel to high temporal sampling rate (e.g. 12-24 
hours sampling time) and large surface coverage. The low 
resolution data mass using in procedures is not too big, therefore 
the necessary computing capacity and the costs are not too high. 
The high resolution data have possibilities in investigations on 
single field levels parallel to low temporal sampling rates (e.g. few 
days ~ few weeks) and not too large surface coverage. The high 
resolution data mass is big, the necessary computing capacity and 
the costs increase. 
ice and ocean research and application; geology; mapping, geodesy, 
production of up-to-date maps; state administration, settlement 
management, land use, monitoring and protection of environment 
etc.; defence, security' and reconnaissance; global changes (Myneni 
et al, 1997); others, as e.g. archaeology. 
In the following let the low resolution (l.r.) data cca. 1 □ 1 km 2 or 
nGn km 2 (n<10) be, the high resolution (h_r.) data cca. kDk m 2 
(l<k<100). 
The main fields of applications: 
a) Crop production: 
□ Land use, acreage determination and mapping □ (mainly h.r. 
data). The accuracy of this application is veiy high, it is tlie same or 
in the most cases better than the accuracy of the classical methods 
(e.g. questionnaire). Besides this the remote sensing method is very' 
fast, can produce early actual land use picture to the administration, 
and cheaper than the classical ones. 
□ Monitoring of plougliing and liarvesting □ (h.r. data). 
□ Monitoring of local stress-effects (disease, flood, drought etc.) □ 
(h.r. and l.r. data too). The Fig. 1. presents the AVHRR greenness 
(GN) □ Ferencz et al, 1993 □ temporal data of the same big com 
field in Hungaiy in normal (1991) and drought (1992,1993) years. 
(DO Y is Day Of the Year.) The effect of drought is remarkably on 
the temporal profiles. 
□ Monitoring of regional stress effects □ (l.r. data). On Fig. 2. 
regional differences of the effect of drought on county-average