ard 
ates 
Photography for stand data collection 
Trials have been made to photograph sample plots from smaller aircraft 
with 70 mm cameras. Scales used were 1:2 000 - 1:4 000. In order to 
sample 150-200 trees and to cover the variability within stands several 
plots per stand have to be photographed. This resulted in a lot of photo- 
graphs to keep track of and a lot of models to treat. An other problem 
is the scale determination of the stereo models. This could however 
be solved with the special radar altimeter /4/ developed in Canada. 
Division into stands and area determinations must however be made in 
photographs in smaller scales. 
Thus it was not found practical to work with 70-mm photography over 
sample plots for stand data collection. Instead conventional photography 
in strips with ordinare mapping cameras (23x23 cm image format) is 
preferred. This offers many advantages: 
+ division into stands and area determination of stands can be made 
+ fewer models to treat 
* ordinary photogrammetric instruments and methods can be applied 
(e.g. aerial triangulation for scale determination if necessary) 
* several stand per model 
Measurements and collection of stand data 
A measuring system for collecting stand data from aerial photographs 
has been developed. The system comprises a Wild Aviograph B8 equipped 
with digitizers. To this is connected an HP 9821 desk-top programmable 
calculator. The calculator is used for numerical relative and absolute 
orientation of the models, for area and slope determination and for 
collection of stand data. Programs for aerial triangulation have also 
been developed. 
Measurement of stand data in the photographs is made from a systematic 
sample of plots. The sample plot is defined by a circular measuring mark 
in the instrument. All trees within a plot are measured. Settings on the 
tops of the trees are made first. Species identification and coordinates 
are stored in the calculator. Measurements are then carried out on the 
ground. For every tree, the calculator computes the location of the 
nearest ground measurement and the tree height is determined. When the 
ground is not visible, a digital terrain model is used to calculate 
the ground elevations. Tree heights and species identification are 
stored on cassette tape. The area of the sample plot is also deter- 
mined, since this varies owing to differences in elevation of the terrain. 
When all the plots in a stand have been measured, the calculator computes 
the stand data. The particulars include the mean height weighed by basal 
area, the mean dbh weighed by basal area, the basal area, and the volume 
and number of trees per hectare. The distribution of the trees 
according to height class and diameter class is also obtained. The 
Standard errors in estimates are also calculated. 
An accuracy study 
A study on the accuracy of the method was made. Measurements were made of 
29 stands with an average area of about 15 hectares. In every stand 
the height of 150-200 trees were measured in the photographs. About 
15 sample plots per stand were needed. Third degree regression functions 
were used to obtain the dbh from the tree heights. Crown diameters were