dir, 
rage 
med 
ich 
pted 
    
As the area of a square Qj is I/N times the size of the total 
inventory area, it can be seen, that every horizontal square 
has an equal probability to be taken into a sample regardless 
on which elevation it may lie. 
As the proportion of horizontal squares on various elevations 
can be estimated without bias, the proportion of other squares 
must, as a sum, become estimated also without bias. The size of 
the unhorizontal squares differ on the photo scale, not only 
through differences in the mean elevation, but also through 
their degree and direction of sloping, and through their position 
in respect of nadir point. By assumptions that these qualities 
are distributed randomly the sampling based on photo can however 
be regarded as unbias also for various classes of unhorizontal 
squares. This is illustrated by Figure 2. 
Fig. 2: The effect of sloping on the size of a square on the 
photo scale 
Pro To UP " 
  
  
  
N 
DA LEVEL oO N 
à ee A” 
\ 90 x 
\ e Alsop 
\ 3. ok 
\ AS \ N 
\ \ N 
N \ X 
X 75 N N 
N A0 \ 
MÉJELD LEVEL = — 
PROJECTION LEVEL 
  
It is supposed in the Figure 2 that the radius hits the square 
in angle 90°-@ for horizontal terrain and 90°-a-8 or 90°-a+8 
for the squares on sloping terrains. The horizontal lenght AB 
represents a side of a square. The lengths A'B' and A'' B''