Function has to be known. This photometric function sure and slope of 1 0°. The right part of field D was wet 
depends on microscattering properties of the individual (approx. 18%), left part was dry (approx. 9%). The ther- 
  
  
  
  
remote surface elements and macrostructure of the soil surface mal measurements of field D were compared to the tem- 
ibution, (Hapke B. W. 1963). Macrostructure can be classified as perature obtained for flat part of field B. The results are 
remote smooth, corrugate or porous. Microscattering properties shown in Tab.9. 
)pogra- can be divided into three general types: forward- , iso- 
to be tropic- or backscattering. Area Field B (flat part) Field D 
To remove negative topographic effect we need to know m. T m To Tc 
the solar illumination conditions on slope surfaces. The 1 14% 16 °C 9% 10.5 °C | 17.9°C 
procedure of removing topographical effect from remote 2 19% 14 °C 18% 87°C | 14.7 °C 
  
  
  
  
  
  
  
  
sensing data usually consists of following stages: 
1) Calculation of slope and exposure of the soil surface Tab. 9. Thermal measurements of field D compared to 
for each pixel of DEM. . | the temperature for the flat part of field B /m - soil moisture, 
2) Evaluation of the solar zenith angle and azimuth for 7 - temperature of field B, T.- temperature of field D (raw data), T.- 
horizontal surface at the moment of registration. For modified temperature of field D (assuming that surfaces are heating 
this purpose it is necessary to know: according to the cosines of incident angle)]. 
e Sun declination and inclination depended on a day 
and an hour of the registration, 
e geographical co-ordinates of the research area: lati- 
It is clearly visible, that applying correction procedure the 
temperature levels (Tc ) of field D are comparable to the 
temperature levels (T) of field B and now could be used 
  
)- tude and longitude of the center of test area. for thermal inertia modelling. 
3) Evaluation of illumination angles of sloping surface 
Saar (solar zenith angle and azimuth) for each pixel on the 
base of data: 
  
slope and exposure of the surface for each pixel, 
e solar zenith angle and azimuth of the horizontal sur- packscattering coefficient 
    
    
  
A | face. (0.8 - 1.7) 
F 4) Generation of the correction coefficient image using EN - 1.0 
known BRDF. ; 
LX 5) Multiplication the raw images by the correction coeffi- 
J cient images. 
 - 30 Many Geographical Information Systems (GIS) allow to 
calculate slope and azimuth from DEM, some of them 
have an option to remove topographical effect using the 
simplest BRDF, basing on the Lambert's formula. The 
Lambert's method was criticized cause over correction of 
the slopes with north exposures, (Smits G.H. et. AI. 
1980). There are also another developed of theoretical 
models (for example: Hapke B.W., 1963, Kimes D.S and 
Kircher J.A. 1981, Cierniewski.J. 1991). 
The test area was topographically diversified with eleva- 
tion differences ranging from 215 to 280 m above see 
level and slopes from 0° to 28°. Corresponding incident Fig. 10. Backscatterig coefficient image. 
angle (solar zenith angle) vary from 30° to 80°. 
For testing the different BRDF it was necessary to prepare 
a special computer program. Two examples of the cor- 
fom images draped on DEM are shown on Lambert's cosfficient 
Mentioned above BRDF models were worked out for visi- 07-23 
ble range of electromagnetic spectrum. There is a ques- t EH - 1.0 
tion: is it possible to use them for thermal range? Does n i er us = 1.0 
ii did radiation depend on registration direction or ? 
not? 
During the thermal inertia modelling for soil moisture 
assessment the problem connected with the calibration of 
the thermal images has also appeared. Temperature of 
the soil surface strongly depends on the exposures and of 
the slope range. The slopes looking North are of course 
  
  
  
  
'OINT - 
  
  
  
  
  
  
  
ield B). cooler of those looking South. In the thermal inertia mod- 
elling for soil moisture evaluation the soil surface tem- 
on level perature for areas of the similar water content should be 
yrizontal on the same level and should not be dependent on the 
- should Surface exposures. 
rved for To check this question some initial measurements were 
al effect made. On the Fig.12 an example of the terrestrial thermal Fig.11. Lambert's. coefficient image. 
tribution Image of filed D is shown. Filed D has a northerly expo- 
285 
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996