, which are 
ontains 
ids TM3 
out module 
display 
switch, and 
with a 
ge from 
tric 
ription 
unted on a 
le. A rectan- 
of a. 
holes on 
+ 45° with 
eter pole 
d 2b. 
dered 
ion board 
nl, rit IS 
us experien- 
ons is 
ending on 
npus. Imme- 
ryatory, 
eration, two 
g the radio- 
esponding 
to the 
- 
3 illustra- 
nd the 
ita sheet. 
>xtreme 
)rresponding 
le is 
, FOr a 
site No.Q 
y diyiding 
"d. 
ibout 50days 
- 20, 1982, 
1 Because 
fa umm — 5 
  
  
IV. EXPERIMENTAL RESULTS' 
  
A casual review of the literature demonstrates an almost overwhelming diversity 
of measurement approaches for the "reflectance" of a scene. The "reflectance" 
referred in this report is actually "reflectance factor" used by Nicodemus 
(1970). It is defined as the ratio of the radiant power, existance, reflected 
by the scene to that which would be reflected by a perfect Lambertian surface. - 
As mentioned in the rpevious section, our calibration board is not quite perfect, 
we take the average of 5 measurements over the calibration board as a reference 
for computing the "reflectance" of the test site at a given looking angle. 
Fig. 4 shows variations of the reflectance ratios R's over the first test area 
as functions of Julian date. As mentioned before, our data began on Date 137, 
‘May 17, 1982. The first three sets of data on the left side were taken between 
August 13 and August 20, which corresponded to the early stage of the second 
crop season. By incorporating these two sets of data tagether, one might get. 
a general view of spectral variations over one full crop season. There are five 
curves, which correspond to five different view angles, for a given spectral 
band. The reflectance ratios for TM3 at first show decreasing, remain little 
changes for about thirty days and finally show increasing during the ripening 
stage. The: reflectance ratios for TM4 and TM5, however, behave quite differently. 
The ratios at first show a little bit decreasing, then, increasing substentially 
and fluctuating, and finally decreasing again. At a given time, the relative 
variations among 5 viewing angles are largest for TM4 and smallest for TM3. 
To see how the temporal spectral variations fit into the rice growing cycle, 
we depict Fig. 5 for this purpose. Fig.5 illustrates a schematic life history 
of 130 day rice crop. At the time of transplanting, the tiller number is so 
small that the paddy field is mainly covered by turbid shallow water and/or 
wet soil. The spectral signature at the earlier "active tillering stage" : 
basically represents by the combination of shallow water and soil. As the rice 
seeding grows and the tiller number becomes large, the green vegetation is 
getting dominate. In the later stage toward maturity, the reflectance from 
brown grains will become competative to that from the green leaves. Basically, 
TM3 is chlorophyl absorptive band, while TM4 and TM5 are chlorophyll reflective | 
band. Thus, the over-all patterns shown in Fig. 4 are in general agreement with 
what we would expect from the life history of rice crop. 
Admittely, there are two substential data gaps in Fig. 4, i.e., during earlier 
vegatative tage and ripening stage. It is worth of comparing our results with 
those taken by a dual-looking ground-based radiometer system (Sung, 1982). The 
system consists of two Exotech Model 100A ERTS Four Band Radiometer. The radio- 
meter are mounted on a hanger, raisded about 4 meters above the surface, with 
one radiometer looking vertical upward and the other looking downward. A self- 
designed data acquisition package is used to record.and preprocess the data 
automatically. The reflectance coefficient is defined as the ratio between the 
data taken downward and upward. The field measurements taken from August 4 to 
October 30, 1981 are plotted in Fig. 6, in conjunction with.our measurements 
for 0° view angle. The dotted lines with the symbols of "Y", "X", and "X". 
correspond to Landsat band 4,5,6 and 7,respectively. The spectral responses 
between band 5 and TM3 and between band 7 and TM4 are quite similar. The fact 
that the variations of band 5 and TM3, as well as those of band 7 and TM3 are 
in general agreement indicates that these two techniques are mutually consistent 
and further strengthen our interpretation. 
To further understand the geometrical and optical properties of the rice canopy 
we plot the "reflectance ratio" as a function of Q, defined as the angle 
between the sunlight toward the target and the sensor toward the target, here 
83 
  
   
  
  
  
  
  
  
  
  
   
  
  
  
   
  
   
  
  
  
   
  
   
  
  
  
   
  
   
   
   
  
  
  
   
  
   
  
   
  
  
  
  
  
   
  
  
  
   
  
  
  
  
  
  
   
  
  
  
  
  
  
  
EAN SEO DE ———— 
— ———— = = 
mum