or by analytical methods, using photo coordinates obtained from a 
comparator and using digital computers to solve the appropriate 
equations. Since the AP-C viewing unit is essentially a stereo compa 
rator, it would be possible to carry out the steps in analytical aerial 
triangulation as an on-line operation. Computer programs for achieving 
this could include very accurate methods for relative orientation and 
absolute orientation based on least - square s solutions using large 
numbers of pass-points or control points. In addition, all of the cor 
rections for air refraction, lens distortion, and other effects could be 
included in these programs. This approach to aerial triangulation 
could realize a considerable reduction in both time and expense. Time 
would be saved because many triangulation computation steps could be 
performed while the measurements were being made on the photographs. 
There would also be a reduction in the time required for rework in case 
of errors because, in most cases, errors in observing points would be 
noticed before the photographs were removed from the viewing instru 
ment, thus allowing the operator to correct any erroneous operation 
immediately. The saving in cost would come about because there would 
be no need for an additional computer or lengthy hand calculations in 
order to perform the analytical triangulation — the AP-C computer could 
perform the entire computation. 
As an example of the use of the AP-C system for analytical aerial 
triangulation consider how various functions could be performed. As 
soon as the last fiducial point on a photograph is located, the location of 
the film principal point, the rotation of the plate, and film-shrinkage 
correction functions may be calculated. As each pass-point or control 
point is located, the image location with respect to the principal point 
of the camera may be determined, considering the plate rotation, film 
shrinkage, and lens distortion. When at least five points on a stereo 
pair of photographs have been located, a first approximation to the 
relative orientation may be found. This may be used to aid the instru 
ment operator in finding conjugate images for succeeding pass-points 
and control points on the pair of photographs. As each additional point 
is measured, the relative orientation may be refined so that, when the 
last point is measured, an accurate relative orientation using all the 
available data may be quickly obtained. The residual y parallaxes at 
all points can then be determined as a check. Points with larger than 
average parallax may immediately be called to the attention of the 
operator for checking.