ESTATE
AS A
m, between calibrations 6 and 7 is undoubtedly the result of the
use of a larger number of points with extrapolated lens distortion
corrections in connection with calibration 73 the number of these
points is given in Table 5.
This conclusion is supported by the results obtained using
calibration 5. Elimination of the 12 points farthest away from the
center of the photograph reduced my from 3.93 to 2.00. Although my
and m, are significantly larger (by approximately 50%) than those
for the other calibrations, X and Y are practically the same. The
effect of the obviously less effective lens distortion correction
in the corner areas of the format is felt primarily as elevation
errors. Table 5 also lists for each calibration the maximum radius
for which the lens distortion was actually determined. These radii
are indicated in Figure 5 from which it is obvious that calibration 5
cannot indicate the steep increase in lens distortion occurring in
the corner areas. The same is true for calibrations 7, 2 and 6
to an increasingly lesser extent. This figure underlines the
necessity to determine the lens distortion as far as possible into
the format corners.
The root mean square positional errors within the groups of points
were computed 2gain for the same calibration as before. They are
3222 T. x e
similar (m, - - 3.0 um, mys X 2.7 um).
4. CONCLUSION
Image quality and lens distortion data obtained from the still
on-going activities of the Working Group on Image Geometry have
been presented in the form in which they were documented by the
working group members. Because of variations in data presentation,
conclusions from data comparisons are general in nature and tentative.
A number of contributions to the program have not yet been
completed and documented. It is hoped that this will be done in time
in order that the report to the 1980 congress will be the final report.
Outstanding are all system calibrations over test fields and the
calibrations carried out with the Itek Artificial Stellar Calibrator.
It is also hoped that all data will become available to permit
processing using the same procedures. Such processing is obviously
required if calibrated focal lengths, reference point locations and
lens distortion values are to be compared with each other.
-22-