3. AIRBORNE TOPOGRAPHIC MAPPER
3.1 Measuring System
NASA has developed several different airborne laser systems,
including the Airborne Oceanographic Lidar (Krabill and Swift
1985) and the Airborne Topographic Mapper (Krabill et al.,
1995a). The latter system was developed for the sole purpose of
topographic mapping, particularly for NASA's Greenland map-
ping program.
The laser scanner of the ATM system covers a 130-200 m wide
swath with a set of overlapping spirals (Figure 1). The trans-
mitter is a pulsed laser that operates in the visible part of the
spectrum. The laser beam is directed along an oval shaped pat-
tern with the help of a nutating mirror. At a nominal operating
altitude of 400 m above ground, the laser spot on the surface
has a diameter of approximately 1 m. In 1991 the scan mirror
was spun at 5 Hz with a laser pulse rate of 800 Hz. The maxi-
mum along-track separation between the laser footprints was 20
m, and the cross-track separation was less than 4 m. In order to
provide higher data density, the laser pulse rate and the scanner
rotation rate have gradually been increased. The present system
uses a laser pulse rate of 5000 Hz with 20 conical scans per
second resulting a very dense data array (Figure 1).
meters across track
meters along track
OFF NADIR: ANGLE (080) terere deis 10.000
AIRCRAET. VELOCITY. (knots)... iei e 330.000
AIRCRAFT ALT ABOVE GROUND (m) .................. ess 400.000
SCAN: RATE (H2)... edttqetinoatratact no dite sonssnsnriasastorssnsaons 20.000
LASER PULSE' RATE (Az) a inn da Lean 5000.000
Figure 2: Scan pattern produced by ATM system (current sys-
tem specifications).
The ATM system is mounted on a P-3 aircraft. The aircraft
location is determined using kinematic GPS technique. The
attitude information is obtained from a ring-laser gyro Inertial
Navigation System. Real-time GPS data are used to provide the
pilot with a visual display of the flight line and the current off-
set from the desired track.
3.2 Calibration and Data Processing
First, the data collected by the individual sensors (laser ranging,
GPS, INS) are processed independently. The various data
streams are synchronized by the GPS time tags. Then they are
combined to provide the 3-D coordinates of the laser footprint
on the surface.
44
3.2.1 Laser Ranging: The round-trip travel time of the laser
pulse between the aircraft and the surface is measured by a
threshold detector. Range determination based on thresholding
depends on the intensity of the received pulse. The increase in
measured slant range with decreasing laser backscattering en-
ergy is often referred to as "range walk". The relationship be-
tween the residual of the true and measured range, and the re-
ceived intensity is established by ground calibration. The
pointing angle of the laser is determined using the rotational
position of the scanning mirror obtained from a shelf mounted
scan azimuth encoder.
3.2.2 Aircraft Attitude: is provided by the INS unit. Data from
three widely separated GPS antennas on the airplane renders
independent estimates of aircraft attitude for monitoring the
INS drift during the flight.
3.2.3 Aircraft Position: is determined by using kinematic GPS
technique, tracking the difference in the GPS dual frequency
carrier-phase-derived ranges from a fixed receiver located over
a precisely known benchmark and a mobile receiver on the
aircraft.
3.2.4 Data Integration: The different data sets are integrated
following a georeferencing scheme similar to the one suggested
in (Lindenberger 1993). The mounting bias between the laser
system and the INS was computed from data sets collected over
flat areas such as the ocean surface in fjords. The following
parameters are available after the data processing: geographic
latitude, longitude, and elevation of the laser footprint
(referred to WGS-84 ellipsoid), scan azimuth, pitch and roll of
the aircraft, and GPS time of the measurements.
3.3 Accuracy Assessment
Principle error sources are related to laser ranging, and to the
determination of aircraft position and attitude. Different tech-
niques are employed for assessing the measurement accuracy,
among them:
e Overflight of runway and apron areas of staging airports
previously surveyed by mobile GPS mounted on a truck.
e Overflight of profiles on the ice sheet previously surveyed
by mobile GPS system mounted on a sledge towed behind
a snow mobile.
Overflight of a profile surveyed by optical leveling;
e Repeat flights and data comparison at “crossing points" of
flight lines.
The results indicate that ice-surface elevations can be reliably
measured by the ATM system to an RMS accuracy of 20 cm,
possibly 10 cm, over baselines of more than seven hundred km
(Krabill et al., 1995a).
3.4 Data Thinning and Blunder Detection
The ATM data sets are very large and not easily manageable.
For example in 1991 one hour flight rendered approximately
2,800,000 data points in a 400 km long swath. Spatial distribu-
tion is quite irregular with redundant data near edges, and small
gaps in the middle of the swath. Outliers are caused by the re-
flection of the laser beam from clouds, ice fog or blowing
snow, or measurement errors. A simple but efficient thinning
scheme reduces the redundancy of the data sets. The recom-
mended processing steps are as follows:
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B1. Vienna 1996
abl
e Bh
e Da
Ice she:
approxi:
selected
scannin,
secutive
1991, m
line. Af
projecti
to deter
remove
ter is us
Using tl
order aj
O, obtz
tion err
follows
vassed :
fjords v
(ice fog
values.
remarka
deviatio
tently b
Greenla
elevatio
Precise
availabl
4. A
The AT
along tt
al., 199.
4. 1 Laı
Contour
the inhe
In order
distribu
lation. I
shows ¢
cated p«
likely s:
4.2 Dig
Data
DEMs
face. B
Fourier
features
towards
from A'
also fac
and oth
