The magnitude of random errors in determining the ZLF can
indirectly be judged by the values of the MSE computed for
various realizations of objects of one and the same class.
These values are governed both by the decree of uniformity
of the objects combined into a class and by the errors in
determining the ZLF, Numerous results of practical measurements
of the ZLF have shown that in most cases the random measure
ment errors do not exceed 10%, For illustration, Table 1 shows
the results of determining the values of ZLF of desert areas
subjected to different anthropogenic effect. The first class
corresponded to the surface in immediate vicinity of the wells,
the second class - to the surfaces located at a distance of
5-7 km from them and the third class - to the background
surface located at a considerable distance from the wells.
In season (September) no plants vegetation was practically
observed.
Table 1
Values of ZLF of desert areas
Class of
obj ects
Number
of
objects
Spectral zones
500-
•560 nm
620-
-710 nm
660-
-870 nm
ME
MSE %
ME
MSE %
ME
MSE %
1
38
0.20
- 7*6
0.31
- 5.9
o.36
- 8.5
2
76
0.18
- 6.0
0.28
- 6.9
0.33
- 4.9
3
76
0.16
- 6.7
0.25
- 8.0
0.32
— 6.8
Thus, the random errors of measurement of the ZLF are
small in magnitude, whereas the systematic errors, particu
larly for low-brightness objects, may be considerable, which
makes difficult comparison of the data obtained by different
survey means at different time.' To disclose and eliminate the
systematic errors it is necessary that artificial or natural
photometric reference standard be available on the terrain.