GV—6 PHOTOGRAMMETRIC ENGINEERING 
If the photogrammetrist is to properly serve his purpose, the photographic 
methods and procedures employed in the analysis of ballistic data should be 
encompassed by him in order that the methods and procedures of photographic 
ballistics and photogrammetry may be exchanged, evaluated, and exploited 
for advancement of the science of photographic measurements. 
There are three approaches to the problem of obtaining accurate data with 
goniometric photographic recorders of static or moving phenomena. 
One approach is to build accuracy into all components of the instrument. 
This results in an extremely costly instrument that is wholly dependent on the 
manufacturer for adjustment. This is the general practice in Europe. 
Another approach is to include provisions for a series of adjustments where, 
with a suitable test range, the errors may be adjusted out from time to time. 
This is time-consuming where observations are being made from day to day in 
large quantities. This is the general practice of surveyors in this country. 
A third approach is to build precision into a minimum number of compo- 
nents, such as the horizontal and vertical axes; then to determine the remaining 
errors by calibration and to enter them as constants in the analytical reduction 
of the coordinates of a point. This is a logical outgrowth of data reduction by 
automatic computers; it, therefore, is the approach of the U. S. Naval Proving 
Ground in system-testing by stars. 
SYSTEM-TESTING By STARS 
A. INTRODUCTION 
Generally, it is highly objectionable to employ one science requiring a special 
set of experiences, formulae, and nomenclature to evaluate a precise complex 
instrument relative to another science that also requires a special set of experi- 
ences, formulae, and nomenclature. The specific reference here is the testing 
of a cinetheodolite for static errors by astronomic methods. Three considera- 
tions are offered to overcome the basic objections: 
(1) Any system developed and component-tested in the laboratory should 
be system-tested in the field if the grossness of the field errors does not camou- 
flage the real errors of the system; 
(2) Observers or operators of the instrument referred to will logically be 
more familiar with astronomic methods than optical bench methods; and 
(3) The operators have an evaluation tool that frees them from a dependence 
on laboratory procedures and equipment. 
The following errors may be determined from a series of exposures made 
at regular intervals in a vertical plane. 
1. Eccentricity of the vertical circle. 
Inclination of the horizontal axis (Ix). 
Inclination of the camera optical axis (7). 
Index error of the vertical circle. 
. Precision of the horizontal axis (X axis). 
The geometry of the observing procedure is illustrated in Figure 1. Assume 
that the exposures are made at a fixed azimuth with vertical angles of 30°, 60°, 
90°, 120°, and 150° in accordance with Figure 1. 
The following errors may be determined from a series of exposures made at 
regular intervals in azimuth, with the camera clamped at some fixed vertical 
angle, say 60°. 
1. Eccentricity of the horizontal circle. 
2. Inclination of the vertical axis (Iz). 
3. Precision of the vertical axis (Z axis). 
Un HB WN