International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
more advanced laser scanners have been built which are capable
of recording more than one pulse. State-of-the-art commercial
laser scanners typically measure first and last pulse; some are
able to measure up to five pulses. Still, the problem is that it is
not always clear how to interpret these measurements for
different targets, particularly if the detection methods for the
determination of the trigger pulses are not known.
Pragmatically, one may for example assume over forested
terrain that the first pulse is associated with the top of canopy
and the last pulse, with some probability, with the forest floor.
However, due to the 3D structure of natural and artificial
objects, the form of the received pulses may be quite complex.
The number and timing of the recorded trigger-pulses are
therefore critically dependent on the employed detection
algorithms. Consequently, it appears to be the logical next step
to employ laser scanners that are able to record the full-
waveform. In fact, first commercial full-waveform laser scanner
systems will become available in the near future.
Another, to a certain degree oppositional trend in laser scanner
scanning, is the design of laser beams with smaller and smaller
beam divergence (tendency to “single mode” signals). With this
sensors the number of multiple returns per emitted pulse will
decrease, due to the fact that a smaller surface patch is
illuminated. Since the acquired information per beam decreases,
classification of the data is only possible in relation to
neighbouring echoes. An interesting aspect for the future
system design may eventually be the combination of narrow
(only one single return with high quality range information) and
wide (recording the full-waveform information) beams in order
to use the advantages of both techniques.
In order to exploit the potential of full-waveform digitising laser
scanners, the physical measurement process must be well
understood (Wagner et al., 2003). In this paper we shortly
review the technical characteristics of laser scanning systems
(Section 2) and discuss basic physical concepts that allow to
understand the way how distributed targets (such as trees or
inclined surfaces) transform the emitted pulse (Section 3). By
taking simple examples, the implications of using different
post-processing algorithms for the determination of trigger-
pulse are demonstrated (Section 4). Finally, section 5 discusses
some of the issues that need to be addressed by future research
and development efforts in order to fully exploit full-waveform
laser scanners.
2. LASER SCANNER SYSTEMS
All commercial airborne laser scanner systems measure the
travelling time of short laser pulses (pulses are typically 5-10 ns
long), but otherwise may vary significantly in their design. For
example, some systems use rotating mirrors as deflection units,
others glass fibres. The laser wavelength is typically in the
range from 0.8 to 1.55 um.
Figure 1 illustrates this distance measurement principle. An
emitted laser scanner pulse (here we use for simplicity a square
pulse) interacts with the earth’s surface. This interaction leads,
if the pulse illuminates a vertically elongated surface target, to a
significant change of the shape of the pulse. The goal of the
distance measurement system is the detection of a previously
defined reference point (based on the emitted signal, e.g. the
raising edge) in the reflected echo. For this task different
detection methods (further details will be presented in section 4)
can be used. In Figure 1 the use of a threshold operator is
demonstrated. In this example two stop pulses are detected. The
use of different detection methods can lead to different results
especially in areas with more than one reflecting element within
one laser spot (see Section 4.2).
4
©
= o
G 3 Emitted impulse
= a
+
LIEF: A
4
2 Reflected impulse
20
oz
se Threshold
e ER
, t, t
4
1. Stop pulse 2. Stop pulse
2%
Bc
2.
Qo uo
t t, t
+ s Digitised reflected impulse
35
29
+ | | |
Oo-c
>
S Ll | m | Lulli i
i t, t
Figure 1. Emitted and received impulse. Commercial systems
estimate the travel time of more than one stop impulses.
Experimental systems digitise the whole waveform of the
received echo with a certain sampling interval.
Current commercial system providers do not offer detailed
information concerning their detection method, so that the end
user has no information about the varying quality of the range
measurements. Therefore the influence of the detection method
on the finally computed models is presently unknown.
The US National Aeronautics and Space Administration
(NASA) has already developed and operated waveform
digitising airborne laser scanners for demonstrating the
potential of this technique for vegetation mapping. For
example, the airborne prototype LVIS (Laser Vegetation
Imaging Sensor) employs a digitiser sampling rate of 500
Msamples per second (Blair et al., 1999). This corresponds to a
range sampling interval of 0.3 m. Hofton and Blair (2002) write
that this sampling interval is sufficient to reconstruct the shape
of the pulse with a vertical resolution of about 0.03 m.
3. PHYSICAL PRINCIPLES
Laser scanning is a direct extension of conventional radar (radio
detection and ranging) techniques to very short wavelengths.
Whether laser scanning is referred to as lidar (light detection
and ranging), laser altimetry, or laser radar, the same basic
principles as in microwave radar technology apply (Jelalian,
1992). As a result, much of the terminology and concepts used
in radar remote sensing can be directly transferred to laser
scanning. In section 3.1 we introduce the radar equation, which
is the fundamental model for describing the measurement
process in terms of sensor and target characteristics. In section
3.2 it is shown that the form of the received pulse can
mathematically be depicted by a convolution between the
emitted pulse and the (effective) scattering cross-section of the
Earth's surface.
Intern
31 R
The in
the ra
target,
where
The ra
of trai
receive
proper
backsc
The b
effecti
taking
reflect
meters
always
(e.g. k
simply
magnit
strong]
orienta
32. P
Model:
differe
Ransor
assume
(becau:
Here w
it is wr
non-til
scatters
range
superp
ranges
expres:
where
and of
the inte
optical
speed «
