7 
A similar automatic feature is provided for absolute orientation 
in which the operator makes observations in the model, enters point 
coordinate data, punches a button and the plotter automatically solves 
and corrects stored constants for all changes to the elements of 
absolute orientation by the method of least squares. 
Its ability to change instrumental and constant settings is similar 
in principal to conventional instruments. The numerical nature of these 
settings allows for convenient application of numerical techniques for 
doing relative and absolute orientation of partial models involving 
high correlation between the elements of orientation. 
6. Systematic Error Correction in the AP-2 
The AP-2 has considerable power and flexibility in the manner in 
which it can take into account various sources of systematic error, 
both known and unknown. As seen from Figure 'the AP-2 console has 
several switches for entering constants to remove the effects of: 
a. Differential film shrinkage 
b. Vehicle motion and IMC errors 
c. Earth curvature and refraction 
In addition, it is also possible to enter constants data into or 
through the GPC computer which are used by the DDA computer to remove 
the effects of lens distortion and to map model coordinates into a 
desired map projection on the plotting table. 
To take care of unknown systematic errors caused by such things as 
errors in relative orientation, incomplete knowledge of lens and camera 
calibration, vehicle motion, etc., whose effects cannot be eliminated 
during absolute orientation, the AP-2 has designed into it what we at 
ACIC refer to as the Black Box capability. It consists of special 
programming in the DDA section to allow warpage of the stereo model in 
the three coordinate directions in accordance with a set of three cubic 
polynominal formulae each containing 9 terms. The coefficients for 
these formulae are developed off line to the plotter and are based upon 
the comparison between stereo model coordinates and given control 
coordinates. Observations are reduced by the least squares method and 
the solved coefficients are taken and entered through the GPC to be 
stored in and used by the DDA computer. The choice of coefficients 
to be solved is completely arbitrary and may be such that the 
correction formulae can have either conformal or non-conformal 
properties as desired. Figure 4 is a graphic representation of how 
this is accomplished.