The effect of the intermediate control zones has already been discussed
(Page 9, para. 3) and it has been shown that had these been omitted the result
would not have been altered by more than about 3ft. at any point. Considering
that the length of the APR bridges between the terminal control of the N-S
strips varied between 36 and 82 miles, the result of this final analysis is highly
encouraging. A total of 194 check readings were made spread over forty-one
overlaps and nine different strips. The mean square error of all readings was
4.2ft.. This included two extreme cases with errors of +15ft. and -16ft.; both
points fell near the centre of overlaps containing at least six other check
points, none of which showed an error of more than 6ft. It seems certain
that the excessive error of the two points was due to misidentification and it
is logical therefore to ignore these two points, in which case the mean square
error of the remaining 192 points is reduced to 3.9ft. and the extreme errors
are +8ft. and -10ft.. All readings were corrected for earth curvature, the
maximum correction being 3ft..
The distribution and frequency of these absolute errors is shown in
Diagram C to be evenly divided between positive and negative values. It is also
to be noted that 8096 of all check points lie within 5ft. of their true value.
The following facts about the survey are worth emphasis:-
1) The APR control flying along the lateral overlaps of the survey
photography was completed in five flying days out of a total of 18
spent on the whole operation.
2) The ground levelling for controlling the APR profiles took 38 man-
weeks compared with the estimate of 170 man-weeks required to
provide normal height control for photogrammetric mapping.
3) The contoured maps of the whole reservoir area were delivered
within three months of completion of photography and APR flying
and one month ahead of schedule.
4) The total cost of the survey with APR control showed a saving of
approximately 23% over the estimated cost of undertaking the work
without APR.
5) 80% of the 192 points checked against ground levelling show no error
greater than 5ft. and, excluding two misidentified points, 100% were
within 10ft..
CONCLUSIONS
It is believed that this is the first APR survey to be undertaken on the
12