ground controls used during these processes are only
accurate to 2cm. The answer is simple. In aerial trian-
gulation or in lidar boresight/calibration there many
variables that are adjusted during the process. These
variables—including exterior orientation parameters,
camera interior parameters, timing, etc.—are con-
sidered adjustable observations with error budget (or
weight) built in, so they are tuned and adjusted during
the process. The nature of the mathematical model-
ing and the least squares we perform during these
processes allow errors in a parameter to change based
on the constraint of that parameter. For example, over
constraining the height of a ground control point in
the solution may push the error in the control to the
adjusted focal length of the camera. The same thing is
valid for easting and northing, as it can be absorbed
by the exterior orientation of the imagery or the image
measurements of the tie/pass points. That is how the
least squares adjustment works—it does not remove
errors but minimizes their effect by redistributing
those errors within the adjusted block.
The previously described reality is forcing us to
reconsider original practices from softcopy or digital
photogrammetry in the 1980s, when the error of
ground survey was ignored while computing product
accuracy.
2) The lack of photogrammetric and surveying knowl-
edge with many data producers, especially the new
UAS-operator-turned-mapper community, leads
them to believe that the residuals in the resulted ae-
rial triangulation or the lidar bore-sighting/calibra-
tion or the fit to the ground controls represent their
final product accuracy. They are not aware that the
fit of aerial triangulation or the lidar bore-sighting/
calibration solutions to the surveyed control or
checkpoints does not directly represent the product
accuracy because it is referenced to the survey da-
tum, which resulted from the inaccuracy in survey-
ing techniques and not to the intended true datum
and coordinates system. Without incorporating
the discrepancies between the true datum and the
survey datum in computing final product accuracy,
product accuracy will be falsely expressed.
3) Creators of mapping standards fell into the same
trap that early mappers fell into by ignoring the sur-
vey error component in calculating product accura-
cy. Users of these standards followed those guides.
By ignoring the error component from the surveying
technique when estimating product accuracy, these
standards contributed to the problem and did not
offer users with a solution.
Based on the previous discussions, the mapping community
urgently needs to embrace the following corrective practices:
1) The mapping community needs to start incorporat-
ing the accuracy of field surveying ground control
points or checkpoints into their product accuracy
computations when reporting final product accu-
racy as illustrated in this article. This will require
negotiating ground control accuracy requirements
with surveyors prior to conducting surveys. Users
will also need to require surveyors to deliver com-
plete survey reports highlighting the accuracy of the
survey. They should consult the ASPRS Positional
Accuracy Standards for Digital Geospatial Data
2
to
understand the required accuracy of ground control
that is needed to meet specific product accuracy.
2) Similar actions need to be considered in the next
version of the ASPRS Positional Accuracy Stan-
dards of Digital Geospatial Data. These standards
need to be amended to introduce the correct way to
compute product accuracy and to provide practical
examples like the ones outlined in this article.
3) Private and public agencies need to mandate that
future product accuracy should be expressed accord-
ing to the new concept introduced in this article. By
not doing so, the stated product accuracy according
to the current practices will be incorrect and mis-
leading.
A
uthor
Dr. Abdullah is Vice President and Chief Scientist at Woolp-
ert, Inc. He is also adjunct professor at Penn State and the
University of Maryland Baltimore County. Dr. Abdullah
is ASPRS fellow and the recipient of the ASPRS Life Time
Achievement Award and the Fairchild Photogrammetric
Award.
1
.
2
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
July 2020
403
