SEMI-REAL TIME MONITORING SYSTEM 
FOR AGRICULTURAL DISASTERS USING NOAA/AVHRR DATA 
G. Saito™*, I. Nagatani®, X. Ssong and S. Ogawa' 
? Graduate School of Agricultural Science, Tohoku University, 1-1,Tsutsumidori Amamiya-machi Aoba-ku, Sendai, 
981-8555, Japan, E-mail genya@bios.tohoku.ac.jp 
p Computer Center for Agriculture, Forestry & Fisheries Research, 2-1-9 Kannondai, Tsukuba, Ibaraki, 305-8601, 
Japan,nagatani@affrc.go.jp i 
* Institute of Geographical Sciences & Natural Resource Researches, Chinese Academy of Sciences, 
Building 917, Datun Road, Anwai, Beijing, 100101,China, songxf@igsnrr.ac.cn 
d National Institute for Rural Engineering, 2-1-6 Kannondai, Tsukuba, Ibaraki, 305-8609, Japan, 
sogawa(@nkk.affrc.go.jp 
Keywords: Remote sensing, Agriculture, Crop, Comparison, Terrestrial, System 
ABSTRACT: 
To use of remote sensing techniques for Agricultural monitoring, it needs semi-real time data acquisition, treatment, and distribution. 
For this reason we are now developing Satellite Image Data Base System (SIDaB). First, I introduce our SIDaB system, and our 
future plane of satellites data acquisition. Using network system such as Asia Pacific Advanced Network (APAN), we get 
NOAA/AVHRR, TERRA&AQUA/MODIS and DMSP/OLS data in semi-real time. 
Next, I mention about vegetation monitoring system. We are also developing Agriculture -monitoring system in East Asia using 
NOAA/AVHRR data. Every ten days maximum Normalized Difference Vegetation Index (NDVI) is compared with past three years 
average of ten days maximum NDVI value. The standard ten-days NDVI images are created using the ten-days data of 1997-1999. 
To reduce the cloud noise, the maximum value of NDVI is used for ten-days composites. When the average value is calculated, the 
minus NDVI pixels are eliminated, since there is the cloud pixels even in ten-days composite. Then, the difference NDVI images 
between the standard ten-days NDVI images and up-to-date ten-days NDVI images are calculated. Using these difference NDVI 
images, it is possible to detect the area of drought damage on agriculture. The minus difference (<-0.1) pixels are listed as drought 
  
risk area in spring and summer. 
1. INTRODUCTION 
In recent years, the satellite, the computer, and network 
technologies have been developed in each year, and these 
technologies have been used popularly. NOAA satellites 
provide a regular, repetitive view of nearly the earth's entire 
surface (Johnson et al. 1993). Many researches have been 
carried out on regional and global land monitoring using 
NOAA/AVHRR data. The AVHRR-based reflectance in the 
visible (VIS) and near-infrared (NIR) wave bands and the 
Normalized Difference Vegetation Index (NDVI) has been 
used for drought monitoring (Kogan 1987, 1994, 1995 and 
1997). 
Agriculture monitoring is important for estimating for crop 
productions and detecting agricultural disaster such as droughts 
and floods. Therefore, the objective of this study is to develop 
and establish the monitoring agriculture system in East and 
South East Asia using NOAA/AVHRR data. The system is 
developed using the changes of up-to-date and normal NDVI 
from AVHRR data. Using this system, the drought effects on 
agriculture in China 2000 year are detected and monitored 
successfully. 
2. USED METHOD, MATERIAL AND SYSTEM 
2.1 Methodology 
The NDVI is very popular parameter from the satellite data. 
Many researchers have used the NDVI for monitoring global 
vegetation and climate studies (Kogan 1990, 1994, 1995 and 
1997; Los et al., 1994; Ramsey at al., 1995; Tucker et. al., 1983 
and 1985; Walker et al., 1998). It is shown that the NDVI is 
useful for these researches. In this research, the changes of up- 
to-date and normal NDVI from NOAA/AVHRR data are used 
for detecting and monitoring drought effects on agriculture. 
Since there are the cloud areas in NOAA/AVHRR data, for 
detecting and monitoring drought cloud free satellite data are 
required. Therefore, to minimize the cloud effect, the largest 
NDVI value of every pixel is used for the daily and ten-days 
images. Although even using this method, we found that there 
are too much cloud cover areas in the daily and weekly images, 
which are difficult to detect drought effects on agriculture using 
these data. On the other hand, monthly data is too long to 
describe the development of vegetation because morphological 
changes and leaf appearances occur at the short period (Illinova 
1975). Therefore, in this research, the detection and monitor of 
drought effects on each ten-days interval are discussed. The 
outline of this research method and the algorithm of this 
research are fig. 1. The standard NDVI images arc created at 
first. Then, the up-to-date NDVI is calculated just after the ten 
days and cloud pixels are masked for elimination. And then, the 
difference NDVI images between the standard and up-to-date 
NDVI images are calculated. Lastly, using the difference NDVI 
image, it is possible to detect the area and intensity of drought 
damage on agriculture. The bigger negative difference pixels 
are listed as drought areas, and the values of difference NDVI 
show the intensities of the drought damages 
2.2 Satellite Image Database System in Agriculture, 
Forestry and Fisheries (SIDaB) 
Every monitoring must be performed on time, and it needs 
semi-real time data for analyzing and old data as standard. For 
this reason we are now developing Satellite Image Database 
System (SIDaB) in Computer Center for Agriculture, Forestry 
and Fisheries Research.  SIDaB gathers raw data 
NOAA/AVHRR data from three receiving stations in Japan one 
664 
Inte 
stat 
DM 
/NC 
Net 
dat: 
The 
Cor 
Res 
Agı 
200 
sate 
(htt 
of 
Shk 
Tha 
(AT 
Syst 
For 
COIT 
soft 
ND 
3.1 : 
First 
the 
peri 
for i 
data 
clou 
surf: 
Gow 
1994 
erup 
1994 
are t 
impc 
cons 
state 
1997 
In o 
imag 
first] 
there 
pixel 
that 
data 
1999 
stand 
3.2 ; 
NDV 
The : 
clouc 
descr 
ten-d 
pixel 
impo: 
these 
area. 
(fig. 
elimi 
The ¢ 
date t