Geography Reference
In-Depth Information
7.2.1.2 Signal registration: echoes
and full waveform
Once the back-scattered energy is collected by the LiDAR's
telescope, two types of signal processing are possible
(Figure 7.1a). The first is a real-time process in which the
received signal extracts the range of targets. Such systems,
called multi-echo or multi-pulse LiDAR, detect signifi-
cant echoes with a threshold method (Thiel and Wehr,
2004). The second possible process is the digitisation and
recording of the entire received waveform. This kind of
system is called full waveform LiDAR. A full waveform
topographic LiDAR typically records the waveform at
a 1-GHz frequency, which is equivalent to one sample
per nanosecond (resolution of 15 cm in air). The target
ranges are then extracted by post-processing the recorded
waveform. A fit of Gaussian functions on the waveform
is traditionally performed, assuming that the signal refec-
tion from the target is Gaussian (Hofton et al., 2000).
Such a process enhances the 15-cm vertical resolution,
allowing the system to fit a Gaussian function between two
samples and to provide more ranging measurements than
multi-pulse systems (Chauve et al., 2009). Furthermore,
full waveform systems can provide the echo amplitude
and width that is useful for characterising the type of
reflective surface (Reitberger et al., 2008).
The ranging measurements are combined with a global
positioning system (GPS) and inertial measurement unit
(IMU) of the aircraft to place the target's position within
a global reference system. Finally, a dense (X,Y,Z) point
cloud coverage is obtained. The points represent the
back-scatter both from objects (buildings, vegetation)
and from bare terrain, and they provide information
about the vertical distribution of targets if the targets are
sufficiently sparse. Indeed, in the case of back-scatter from
a building, the laser will be completely reflected from the
building surface, but in the case of sparse vegetation (see
Figure 7.1a), the laser pulse will be reflected three times.
The first echo will give the maximum elevation of the
vegetated surface while the last echo will provide the
elevation of the bare Earth. Figure 7.1b shows that the last
returns are better able to provide bare-Earth elevations
under sparse object cover and to provide information
on the internal structure of the target. Nevertheless, the
first returns better describe the top of the surface, which
corresponds to vegetation canopies or roofs.
7.2 Ranging airborne LiDAR physics
7.2.1 LiDARfor emergent terrestrial surfaces
When considering emergent terrestrial surfaces, both the
ground and vegetation have a high reflectance ratio in
the near-infrared range (NIR), 700 nm-1,400 nm (Caloz
and Collet, 1992). Vegetation also has a high transmission
ratio in that range, which facilitates the travel of an NIR
laser to the ground through the vegetation. As a result,
LiDAR, using an NIR wavelength called topographic
LiDAR, is able to simultaneously gather information on
both the ground and vegetation.
The most commonly used infrared wavelength is 1,064
nm, which is the characteristic wavelength of frequency-
doubled Nd:YAG (neodymium-doped yttrium alu-
minum garnet; Nd:Y
3
Al
5
O
12
) diode-pumped, crystalline
lasers. Nd:YAG lasers are preferred as they induce greater
power efficiency per transmitted pulse that promotes
a high signal-to-noise ratio to ensure a full waveform.
However, other lasers are also used in topographic LiDAR
(Jutzi and Stilla, 2006), such as erbium fiber amplified
lasers, which provide a 1,550-nm wavelength with
more powerful emitted pulses (Samson and Torruellas,
2005).
7.2.1.1 Physical equations
LiDAR for terrestrial surface measurements emits pseudo-
Gaussian laser pulses and receives the sum of the pulses'
energies reflected by each target the laser beam reaches.
Consequently, the received waveform is the sum of the
response functions of the different targets, convolved
by the transmitted signal and by the receiver impulse
function.
The received power
P
(
t
) can be written as (Wagner
et al., 2006) (equation 7.1):
N
S
fov
P
T
(
t
)
∗ σ
i
(
t
)
∗ Γ
(
t
)
P
(
t
)
=
(7.1)
2
R
i
β
t
π
i
=
1
where
Γ
(
t
) is the receiver impulse function,
σ
i
(
t
)is
the back-scatter cross-section that combines the target
parameters,
P
T
(
t
) is the transmitted signal,
S
fov
is the area
of the receiver optics,
β
t
is the transmitter beam width,
R
i
is the target range, and N is the number of targets.
Search WWH ::
Custom Search