International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004
(iMARG iNA V-ROH) from 30-180 seconds. These results have
been obtained using the University of Calgary KINGSPADTM
GPS/INS integration software (www.kingspad.com). Further
improvements in position and azimuth accuracy can be
achieved if an odometer is added for length control and a two-
antenna GPS for azimuth control. An interesting example on the
use of a road vehicle in an extreme situation, the urban canyon
environment of central Tokyo is given in Scherzinger (2002).
Although adequate GPS coverage was denied for about 50% of
the time, a position accuracy of better than25 cm was
maintained for about 90% of the survey.
KF (m) Vehicle
motion
30 sec 0.08 Straight line
60 sec | 0.09 Circle
30 Sec 0.13 Curve
60 Sec 0.06 Static
120 Sec 0.12 Static
180 Sec 0.3 Straight line
Table 3: Accuracy of the IMAR® iNAV-RQH INS system in
stand-alone mode.
The third example given here is a backpack MMS developed at
the University of Calgary. The backpack MMS competes in
both accuracy and initial cost with current methods of GIS data
collection, while offering increases in data collection efficiency
and flexibility that only an MMS can provide. The backpack
MMS uses a Leica Digital Magnetic Compass (DMC) for
attitude determination, a single Kodak DC260, and a Novatel
OEMA receiver. The system's operational steps are essentially
the same as any MMS in the sense that the DMC and the GPS
provides the direct georeferencing information for the cameras.
The only difference is that because of the low accuracy of the
DMC derived attitude, the system is augmented with a Bundle
Adjustment software, see Ellum (2001). In this case, the direct
georeferencing parameters are used as apriori information on
the exterior orientation parameters. The absolute accuracies of
the Backpack MMS in a variety of configurations are shown
Table 4. From the tables, it can be seen that with as few as five
image point measurements at a 20m camera-to-object distance
it is possible to achieve accuracies that satisfy many mapping
applications, for more details see ibid.
The systems presented so far are working in post-mission mode
of operation. In this mode of operation, the data is collected in
the vehicle (van, airplane, or ship) and processed off-site in
order to extract the information of interest. Due to the post-
mission mode of operation, very high accuracy, in position (x
0.1 m "RMSE") and attitude (= 0.02 degrees "RMSE") can be
achieved. This is accomplished by using the precise GPS carrier
phase in Differential mode (DGPS uses two GPS receivers; one
in static mode over known control point and the second on the
vehicle. By tightly coupling DGPS and INS data through
Kalman filtering the above accuracies can be achieved.
Number of Horizontal Vertical
Ysmnana Daimt
RM Ma RM
S. X S
(m) (m) (m)
3 Images 3
5 Image Points | 0.13 | 0.10 | 0.10 | 0.06
Max
(m)
image 13 010012) 007
Points
| 6 Images |
5 Image Points | 0.06 | 0.05 | 0.14 | 0.11
10 Image
Points 0.07.1 0.05.1:0.17 1. 0.14
Table 4: Backpack MMS System Absolute Accuracy (20m
camera-to-object distance)
Although in most remote sensing applications, there is no need
for real-time processing of the data, there are some emerging
applications, specifically forest fire fighting; in which the
requirement for real-time mapping is more important than the
achievement of highest possible accuracy. One of the main
problems in combating forest fires is monitoring the time
history of the fire. Understanding the size, location, and speed
of advance of the fire front is critical to the optimal allocation
of fire-fighting resources and the maintenance of the fire crew
safety. Investigations of major wild-land fire accidents
involving loss of life, often indicate that the crews became
imperiled because of insufficient or untimely information about
the location and speed of the fire advance.
The F? system, being developed at the U of C, integrates
imaging sensors (Thermal InfraRed “TIR” Cameras) with real
time navigation technologies (Wide Area Differential GPS
“WADGPS” and low cost INS). The system is very useful in
reporting the exact situation of fires, assisting the Forest Fire
Information Systems (FFIS) in accurately assessing the fire and
precisely directing water-bombers and fire-fighting crews. The
use of infrared/thermal cameras, which sense the heat emitted in
the form of infrared radiation, will enable early detection and
location of forest fires in reduced visibility due to haze, smoke
or darkness. Recent system testing over controlled fire pits of
known coordinates indicates that the system’s real-time
positional accuracy in identifying hotspots is about 6m RMS
(Wright and El-Sheimy, 2003), when using single receiver
pseudo-ranges..
Kinematic GPS controlled aerial photogrammetry and direct
geo-referencing using DGPS/INS have become mature
technologies in both the scientific and commercial mapping
communities. Virtually all airborne mapping systems now
integrate a GPS receiver and INS unit with their camera.
Unfortunately, on the software side, the integration of GPS or
GPS/INS and photogrammetry is not as close. Typically, the
GPS data is included in the photogrammetric bundle adjustment
as processed positions only (Schmitz et al, 2001; Mikhail et al.,
2001) and similarly the GPS/INS data are used for direct
georeferencing. In effect, the processing engines of GPS (or
GPS/INS) and photogrammetry operate largely in isolation.
This implementation has obvious benefits in terms of
simplicity. However, a more fundamental fusion of the GPS and
Wi
ve
ac
ur
an
va
Ov
dr
Fu
co
fig
La
Bi
