In the bundle adjustment of the block of 28 photos the same 
subset of error coefficients was exercised as in the Vermont Project. 
Surprisingly, the resulting systematic corrections turned out to be very 
similar to those obtained in the Vermont Project. In particular, a large 
correction for radial distortion was obtained, as is clear from Figure 9 
which presents a vector plot over the format of the recovered systematic 
corrections. This demolished the 'thermal gradient' hypothesis advanced 
earlier, for very nearly the same radial distortion function was recovered 
in the Atlanta Project, even though the camera was subject to much more 
moderate variations of temperature because of the lower flying altitude. 
Accordingly, it became necessary to seek another explanation, a topic that 
will be addressed shortly. 
After completion of the photogrammetric reduction DBA sent the 
coordinates of all of the offset targets to the Georgia DOT. Also sent 
were the coordinates of all installed monuments as computed by DBA from 
the necessary offset measurements supplied by the Public Works Department 
of the City of Atlanta. After receiving the final coordinates from DBA, 
the Georgia DOT made a selection of the points to be field checked and 
conducted the necessary surveys. A total of 18 horizontal checks were made 
by means of a first order traverse. These corresponded to six triplets, 
each consisting of two painted targets and the associated monument. In 
addition, a total of nine vertical checks (i.e., three triplets) were made 
  
— 72 
TE 
We 
fe 
TU 
INA 
m as nmt 
— 
— 
— 
— 
——— 
ult NS \ 1 \ s s \ \ l "Aa 
BN oi olo Fs ot a ET NT eoa Ie 
"sS so &absqu8 aide Purser i SN eli thes our [TE 
Ww I Ls - : 4 : > ' , 2 Ar 
Od ee — = 4 4 ‘ - -— = oerttaa A 
0 —= TT ~~ 7 ? 4 i Y ~ 
— 
NO US MM 
S ee 
EN 
— 
  
  
OA 
T b v4 
80 100 110mm 
Ds 
edt—— 
s 
BA mm 
  
' 0^ 20 40 
FIGURE 9. Vectors of total systematic corrections recovered 
by bundle adjustment with self-calibration on Atlanta Project. 
2.