910
December 2016
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
Figure 1: Control Point distribution with Used and Unused
Points.
S
tudy
A
rea
and
G
round
C
ontrol
Study Area
In order to comply fully with the Section 333 rules in place
at the time, Keystone used a farm in Eastern Pennsylvania
as a test location. AGA farms is located in Perkasie, PA, ap-
proximately 26 miles from Keystone’s office, 15 minutes by
air or 34 miles by car, one hour in morning traffic. The entire
farm is approximately 98 acres with the specific area of in-
terest within the farm being 10.6 acres. The crops grown on
the farm include pumpkins, corn and Christmas trees with
terrain consisting of undulating bare crop land and some tree
cover.
Ground Control
Before the flight day was scheduled, a team from Keystone
placed 18 control points, in a regular pattern, throughout the
test area. These control points consisted of alumi-
num strips arranged in an X pattern and paint-
ed white with a ground station GPS unit used
to record each location. The locations were
processed via the Online Positioning
User Service (OPUS) with a hori-
zontal and vertical accuracy bet-
ter than 3cm. Keystone does
not perform land survey-
ing work as part of
its business and so
repurposed its
base station
e q u i p -
ment for
the task
in order to
minimize
the
al-
ready sub-
stantial cost
of this operation.
After using the control
within the project, it was
determined that three points were incorrect and these were
removed from the solution, however this still left 1.4 control
points per acre over the area of interest. This excessive con-
trol layout was designed to allow the processors to experi-
ment with varying amounts and locations of control while
still having a significant amount left for check points.
E
quipment
and
D
ata
A
cquisition
Equipment
Keystone used its fixed wing, single engine Cessna 210 to
acquire the airborne imagery using an UltraCam Falcon
Prime (now called Falcon Mark II). The UCFp is a large for-
mat camera from Vexcel Corp. that captures RGB and NIR
simultaneously with high accuracy GPS/IMU positioning.
The Falcon has a focal length of 100mm, a pixel size of 5
microns and a frame size of 17,310 x 11,310 pixels. For the
lidar acquisition, Keystone’s fleet of Cessna 310’s was tapped
and the aircraft was fitted with a Teledyne Optech Galaxy
airborne laser terrain mapper (ALTM). This sensor has a dy-
namic field of view (0-60 degrees) and multi-pulse setting.
The Galaxy also has the ability to record up to 8 returns and
intensity measurements per pulse.
For the unmanned systems, two platforms were used for
data acquisition. The Altavian Nova F6500 is a fixed wing
aircraft with a wing span of 9 ft. The aircraft weighs 15lbs
with its payload and battery with a time aloft of over one
hour. The aircraft has a max altitude of 10,000 ft. and a
max speed of 70 mph. The F6500 is equipped with a
Canon EOS Rebel SL1 camera with a 20mm fo-
cal length lens. It is a 22- megapixel cam-
era with a 51,843 X 3,456 pixel frame.
The camera has been modified
for acquisition of both RGB and
CIR (though not simultane-
ously). By changing filters and
white balance settings, the cam-
era can be set to capture RGB in
one flight and CIR in the next.
One of Keystone’s rotorcraft
UAS was used on the project,
the SteadiDrone Mavrik X8. The
Mavrik is an octoquad multi-
rotor UAS. Octoquad is a term
meaning that the UAS has four
arms and eight motors with two
motors on each arm (with one
motor on top and one motor be-
low). It has a max takeoff weight
of 18kg, max speed of 20 m/s
with a time aloft of approximate-
ly 10 minutes with a large payload and closer to 20 min with
a lighter payload. The Sony A7R camera was mounted to the
UAV, which is a substantial payload of 2.5 kg, but is an ex-
cellent mapping camera. It is a 36-megapixel camera with a
7,360 x 4,912 - pixel frame and a Zeiss 35mm f/2.8 lens. The
combination of large footprint and quality optics makes this a
good choice for photogrammetry despite the short flight times.
