The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B5. Beijing 2008
Figure 8. Position Differences, 300 km Network
2.1.2 Simulated Satellite Outages Due to Large Roll
Angles: In order to study the effect of satellite outages that
might occur during a high banked turn, two types of outages
were simulated. This was done by turning off the raw GPS
Observations for some of the satellites in the processing at the
centre of the trajectory. The first case reduced the number of
satellites being tracked to only 2 during a period of 30 seconds,
while the second case reduced the number to 2 for 75 seconds.
In each case the SBET was re-generated using the 60 km
SmartBase network and then differenced with the POSPac V4.4
reference solution that did not have any satellite outages.
Figure 9. Position Differences 30 sec Outage
Figure 9 shows the position differences for the simulated outage
of 30 seconds, while Figure 10 shows the differences for the 75
second outage. In both cases, once the solution has converged,
the position differences remain below 10 cm, and are virtually
unaffected before and after the outages.
Figure 10. Position Differences 75 sec Outage
2.2 Applanix DSS Case Study
As part of the USGS certified manufacturing process for the
Applanix DSS, a directly georeferenced medium format camera
system, each DSS undergoes a final acceptance test in the form
of a flight test over a geometric test range. The acceptance test
involves validating the calibration of the DSS and the accuracy
of the embedded POS AV 410 solution against a series of
ground control points.
As of October 2007, Applanix began using POSPac MMS and
flying sharp turns on all DSS acceptance tests. This section
presents a summary of the geometric analysis results for 8 of
the test flights, plus an analysis on the associated savings
attained.
2.2.1 Accuracy Results, Sharp Turns: To evaluate the DSS
a small block of photos consisting of 12 flight lines is flown.
Ten of the flight lines are flown at a Ground Sample Distance
(GSD) of 15 cm, while 2 of the flight lines are at flown at either
18 cm or 25 cm GSD, depending upon the focal length (60 mm
vs 40 mm). Each line contains 6 photos, and they are flown in a
cross pattern for strong geometry. Endlap and sidelap is 60%
and 40% respectively. For this analysis only the 15 cm GSD
flight lines were used in order to allow a direct comparison on
accuracy between flights. Geometric accuracy was assessed by
running the block of photos in the POSPac Calibration and
Quality Control (CalQC) module along with the Exterior
Orientation computed from the POS AV 410. The CalQC
automatically generates a set of tie points from the imagery,
then runs a block adjustment holding the EO, camera model and
boresight angles fixed to see how everything “fits” against
surveyed Ground Control Points (GCP). The RMS accuracy
against the GCP gives a direct indication on the quality of the
EO from the POS AV.
A total of 2 to 8 GCP’s were used to evaluate each flight,
depending upon what could be identified in the imagery due to
weather conditions (ie snow coverage). The average number of
measurements per GCP ranged from 10 to 15 (although only 2
GCP is not ideal, since the imagery was flown with such high
overlap the same GCP could be viewed in multiple images,
hence the high number of measurements).