Orientation Element Différences 
to Investigate Error Causes in Aerial Triangulation 
Robert B. Forrest 
Photogrammetrist 
Cornell University, Arecibo Ionospheric Observatory 
Arecibo, Puerto Rico 
For many years, photogrammetrists have been subjected to dozens of 
papers on theories of errors and error propagation in aerial triangulation. Some 
of these hypotheses are in direct opposition - yet they flourish in the absence 
of empirical evidence that would confirm or refute the opposing theories. 
The research that I am discussing today is empirical rather than theoreti 
cal. I did this work because - after reading some theoretical studies - I was 
curious to see some real orientation errors for an actual strip triangulation. 
After determining these orientation errors, I made graphs of the errors in 
successive photographs to show the actual error propagation of orientation 
errors for a real triangulation. 
I repeated this determination of orientation errors for several triangulations 
to find to what extent the error pattern of each orientation element is a function 
of the triangulation process. The error graphs of each orientation element for 
the several triangulations provide some information on the actual propagation 
of orientation element errors. In addition to these graphs, I performed some 
further work which shows the importance of relative orientation errors in the 
final residual coordinate errors of the triangulated points. 
I will now describe in more detail the materials and procedure that I used 
in this empirical work. 
Materials — A 37 exposure strip of Wild RC-7 glass plate photography (Aviogon 
objective, f=100 mm., 14 x 14 cm. format) was used for this work. The strip 
was taken in October, 1954, as part of a research project of the I.S.P. Stato- 
scope data was obtained during the photographic mission. The flying height 
was 6,100 m. above terrain, and the strip length is 106 km. Approximate photo 
graphic scale thus is 1/60,000 and the base-height ratio is 0.5. Relative 
relief ranged up to 13 per cent of the flying height above mean terrain. Over 
20Ô geodetic control points are available for this strip, an average of 12 per 
photograph. 
Triangulations - Three instrumental triangulations of the entire strip were 
performed, triangulations 1 and 2 by aerial polygons, and triangulation 3 
(performed in 1959 by another operator) by aerial levelling. Triangulation 4 was 
performed for the first nine models by aerial polygons. Triangulation 5 was 
started by changing photograph 10 to the left projector and setting in the orien 
tation elements determined for photograph 10 during triangulation 4. The 
optical train was adjusted to obtain an erect stereoscopic model, but With left 
and right reversed. Triangulation 6 was performed for photographs 24 through 
37, starting with the elements of photograph 24 set to the orientation elements 
determined in triangulation 2. Triangulations 1 through 6 were performed at 
1/15,000 using the Wild A7 Autograph of the Ohio State University. Later an 
analytical triangulation also was performed running from photographs 7 through