408
June 2019
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
such as orthoimagery and/or 2D vector maps with a horizon-
tal accuracy class of 30 cm, your ground control points need
to be surveyed with:
–
–
Horizontal accuracy of RMSE
x
or RMSE
y
= 7.5 cm
–
–
Vertical accuracy of RMSE
z
= 15 cm
However, if you plan to produce any elevation data, such as
point clouds, contours or 3D vector maps, your ground con-
trol points need to be surveyed with:
–
–
Horizontal accuracy of RMSE
x
or RMSE
y
= 7.5 cm
–
–
Vertical accuracy of RMSE
z
= 7.5 cm
Question 2—
In section 7.9 of the standards, when it
talks about checkpoints being three times the accuracy
of the geospatial data set being tested, does that mean
the GCP accuracy will be three times more accurate
than the desired/intended accuracy class?
Answer:
Yes, it does. The standards in section 7.9 state the
following:
“The independent source of higher accuracy for checkpoints
shall be at least three times more accurate than the required
accuracy of the geospatial data set being tested.”
Therefore,
if you are testing products that must meet a vertical accura-
cy class of 10 cm, your checkpoints should be surveyed to a
vertical accuracy of RMSE
z = 3.33 cm.
Question 3—
Do you use Table D.1 to calculate all the
statistics and then use the results to determine the
ASPRS accuracy class? Is that the typical workflow? Is
there a sample report you can supply to me?
Answer:
Table D.1 in the standards represents the meth-
odology reported in the “National Standard for Spatial Data
Accuracy (NSSDA)” testing guidelines. While you can use
any organization to compute the statistical terms needed to
calculate the accuracy figures (according to ASPRS stan-
dards), Table D.1 is particularly helpful.
Question 4—
Are there guidelines on what you should
aim for regarding the additional statistics discussed
in the standards, such as max, min, skew, kurtosis and
mean absolute?
Answer:
No, not at this moment. However, textbooks and
manuals on statistics can be consulted to derive desirable
values based on the limits of RMSE and bias set by the stan-
dards. I will add your question to the list of issues catalogued
for future enhancements to the standards.
Question 5—
Can you clarify what it means when you
can state, ‘tested to meet’ versus ‘produced to meet?’
Answer:
Formal testing statements were provided to users
in section 7.12 of the standards. There are two types of state-
ments for reporting product accuracy:
1)
’PRODUCED TO MEET’
: This statement is usually
provided by data producers or providers when accuracy
is not verified by an independent set of checkpoints.
The producers report their achieved accuracy based on
confidence in their workflows and by the data fit to the
control they used to calibrate the products.
2)
’TESTED TO MEET’:
This statement is usually pro-
vided by data users or their consultants when accuracy
is verified using an independent set of checkpoints. For
product accuracy to be independently validated ac-
cording to ASPRS standards, the test must satisfy the
following conditions:
a.
Independent checkpoints are ground control points
that are not used in the calibration process during
product generation. Check points can also be derived
from existing datasets with known accuracy.
b.
Checkpoints should be more accurate than the tested
products. According to the standards, checkpoints
should be at least three times more accurate than the
tested product.
c.
To make it a valid statistical sample, regardless
of the project size, there should be at least 20
well-distributed checkpoints used in the accuracy
assessment.
Question 6—
Can you direct me to a document
regarding planning and best practice guidelines??
Answer:
Unfortunately, no such document exists at ASPRS
for UAS-related activities. I co-instruct several workshops
during the ASPRS annual conferences that you may find
useful. In those workshops, we explore best practices based
on past project experience and technical guidelines devel-
oped over the years from our standard mapping practices.
However, I am working with the ASPRS UAS subcommittee
to encourage them to develop such a document that we can
include as an addendum in a future version of the standards.
(to be continued)
**Dr. Abdullah is S Chief Scientist and Senior Associate at Woolpert,
Inc. He is ASPRS fellow and the 2010 recipient of the ASPRS
Photogrammetric Fairchild Award.
The contents of this column reflect the views of the author,
who is responsible for the facts and accuracy of the data pre-
sented herein. The contents do not necessarily reflect the offi-
cial views or policies of the American Society for Photogram-
metry and Remote Sensing and/or Woolpert, Inc.
“’PRODUCED TO MEET’ statement is usually
provided by data producers or providers when
accuracy is not verified by an independent set of
checkpoints.”
“’TESTED TO MEET’ statement is usually
provided by data users or their consultants when
accuracy is verified using an independent set of
checkpoints.”
