139 
'Unt 
elf 
anting 
endent 
3 pro- 
t amount 
no st 
Geese, 
3 to be 
this 
/eloped 
lie of 
^ed by 
ible. 
’a 
ï two 
>rovides 
i long 
.arge 
* the 
Ideation 
and of 
cro- 
ial 
the 
can 
set 
aper- 
one 
effect 
that at 
During our Investigations all the three different 
apertures were used depending on the scale of aerial 
photographs. Some aerial photographs were scanned using 
both 50 and 100 micrometer aperture for the purpose of 
comparison. 
The scanning densitometer is connected to a magnetic 
tape unit which records the densities of images in binary 
form IBM compatible. Data recorded on magnetic tapes are 
retrieved using an IBM 370 electronic computer. 
The conversion of data to visible output requires the 
use of computer programs. An optional program determines 
the number of records and record lengths of a magnetic tape. 
Another program converts binary data on a 9-track magnetic 
tape, originally produced by the scanner, to IBM 370 com 
patible binary halfwords on tape. Data recorded on this 
tape can be used for matrix reduction (automatic count). 
If print out of these data is required, a third program 
has to be used. 
Pattern Recognition 
The procedure of automatic count is a difficult one. 
First, density patterns of all possible images has to be 
established. Then every uniform pattern detected on the 
photos must be compared to the pre-established patterns, 
sorted accordingly, and counted by the computer. 
DATA ANALYSIS AND RESULTS 
Large scale experimental aerial photography for Greater 
Snow Geese, Eider ducks and diving ducks was carried out in 
Eastern Canada and on the Bylot Island, N.W.T. Using the 
obtained photographs color infrared transparencies of eider 
ducks, and black and white pictures of Greater Snow Geese 
and divers were digitized and density patterns were esta 
blished. Preliminary result of this investigation was 
reported in 1972 (1). Density patterns of several classes 
within each species were recognized. However, actual class 
identification was not possible due to lack of ground truths. 
Since sufficient financial support was not available for de 
tailed studies in subsequent years our investigation was 
confined to the counting of the Greater Snow Geese as this 
species represents simple image patterns on the photographs.