ten use
nber, re-
of photo
measure-
to some
e factors
s height,
fferences
the first
:'ondition
istic pat-
he forest
re visible
occurred.
burns or
orest in-
-size con-
ng to the
hey were
on aerial
systems.
the field,
pretation
1 only be
ut if the
1 photos,
ams and
retations
- of days
tof their
of forest
otos.
rime im-
en meas-
iscussed.
lassifica-
f contact
Then the
asses are
ral office
oth field
IC ENGI-
d on the
PHOTO INTERPRETATION IN FOREST INVENTORIES
Fic. 5. Transparent templet orie
Only a fraction of the total area is
of accuracy for most forest surveys, and can be secured far c
procedures.
nted over each contact print marks the plots to be classified.
actually classified, but the resulting data will meet the standard
heaper than by normal mapping
Sampling procedure.—^ simpler, faster, and less expensive method of
measuring forest areas has been developed. It makes use of sampling pro-
cedures and statistical methods. A transparent templet (Figure 5) with a
number of sample plots evenly distributed over an are:
the aerial photos used, is oriented over each contact print that is within
of
a equal to the net area
the tract to be measured.
Each plot is examined under stereo and classified according to the pre-
scribed scheme. If 20 per cent of the sample plots are in a given class, it is
assumed that 20 per cent of the total area is in that class. By applying pro-
portions obtained in this way to an accure
itely computed tract acreage, the
area in each classification can be obtained. Accuracy of course depends
largely upon the number of sample plots classified.
Photo interpreters find the procedure easy to learn and use because they
can concentra
Although only
generally results in inventories as
tained by normal mapping procedures.
te their efforts on plots small enough to fall into a single class.
a fraction of the total area is actually classified, the method
accurate as and cheaper than those ob-
