Mr. C.U. Thorsell Kjell Degerstedt, Per Jakobsson and Torbjórn Olofsson in Lantmáteristyrelsens mätningstekniska
avdelning have all taken part in the detailed analysis of data. Mr, Thorsell wrote the programs for our computer.
Mr. Degerstedt advised us as to theoretical and applied statistics. Mr. Jakobsson prepared all data and performed
special investigations, assisted by Mr. Olofsson. Professor Edgar Dauphin from Kingston, Canada checked the trans-
lation.
We thank them all for their valuable contributions to the Experiment Pecny.
Sven G. Móller organized the Experiment, performed the analysis of error and prepared this report.
1. GENERAL POINTS OF VIEW,
We will study the accuracy of good every -day international photogrammetry, the real accuracy, which would normally E e
result from a group of instruments, handled by a group of operators using many methods of relative and absolute orien- e
tation. The intention is not to study scientific results from toptrimmed instruments and operators, using sophisticated
methods, We concentrate upon the requirement of the consumer of photogrammetric products.
The errors will be stated as the differences (x, y, z) between geodetic coordinates (X, Y, Z) of comparison points within
the test -field Pecny^ and the corresponding photogrammetric coordinates (x' , y* , z' ) from the large group of parti -
cipants in the Experiment. About 50,000 coordinate differences are available. They constitute the error population of
the Experiment, It will be considered as the international photogrammetric error population. Errors from one single model
must always be considered as one sample of many possible samples. Inference from sample to population will be tested
with the aid of modern statistical methods.
The total photogrammetric error can be split into gro ups of partial errors, at least theoretically. It is usual to group
them according to the sources of error and at the same time in the consecutive order in which they occur in the proce- eo r
dure of photogrammetric work. The sources of error, arranged in this order, are
Constant Stochastic
geodetic surveying diapositive
signalization stereoinstrument
curvature of earth relative orientation
refraction absolute orientation
light conditions stereooperator
aerial camera point identification
aeroplane photogrammetric measurement
exposure conditions coordinate registration, etc,
negative - also - negative
photographic procedure - also - photographic proc edure
film shrinkage - also - film shrinkage
Each group can be subdivided into partial errors, They will, however, be very small and of little importance. It would
surely be possible to find about 100 or more partial errors,
