onboard an
UAV
. The sensor was a pushbroom scanner with
linear
CCD
arrays operating in the spectral range of 350 nm -
1030 nm with 128 bands and 5 nm of spectral resolution and
1,024 pixels per line. A field campaign was conducted over
the Baotou test site in China deploying portable reference
reflectance targets.
A radiometric calibration based on the vicarious method
was proposed in Pozo
et al.
(2014) with the aim of adjusting the
involved parameters to better collect the information provided
by the sensor. The multispectral sensor is a
CMOS
-based with
six channels, weight of 1025 g, and geometric resolution of
1280 × 1024 pixels in size. This system was installed onboard
an octo-copter with weight of 2,420 g including the battery.
Imaging spectrometers were designed with high technolo-
gy to work as hyperspectral devices onboard the
UAVs
(Hruska
et al.
, 2012). The integration of these devices with
IMU
and
GPS
allows obtaining direct imaging georeferencing after im-
age processing. Additional efforts for spectral calibration of
hyperspectral data observed from a hyperspectrometer have
been reported in Liu
et al.
(2014).
The ability of
UAVs
to fly at low altitudes, equipped with
specific technologies, allows the acquisition of images with
both, ultra-high spatial and spectral resolutions. Lucieer
et al.
(2014b) described the design and operability of a new hyper-
spectral
UAS
(HyperUAS), a multi-rotor helicopter carrying
a pushbroom spectroradiometer in conjunction with a dual
frequency
GPS
and an
IMU
. The HyperUAS prototype acquires
hyperspectral images with 324 spectral bands and 2 to 5 cm
spatial resolutions after spectral and radiometric calibration
and atmospheric correction. Burkart
et al.
(2014) developed a
hyperspectral measurement system for
UAVs
, operating in the
spectral range of 350 nm - 800 nm, based on the Ocean Optics
STS microspectrometer with a weight of 216 g.
Suomalainen
et al.
(2014) designed a lightweight hyper-
spectral system, with off-the-shelf components, for rotor-based
UAVs
with weight of 2 kg. It consists of three elements: a push-
broom spectrometer with spectral range of 400 nm - 950 nm
and spectral resolution of 9 nm, a photogrammetric camera,
and a
GPS
/Inertial Navigation System. Geometric and radiomet-
ric procedures are designed for
DSM
production in agriculture.
Zarco-Tejada and Berni (2012) equipped a fixed-wing
vehicle with a micro hyperspectral imaging sensor for vegeta-
tion monitoring. These authors and co-workers in the research
group QuantaLab-
IAS
-
CSIC
(2015) have also used multispectral
and hyperspectral sensors for different purposes, mainly in
agriculture and forestry, as described later. Plate 3 displays a
multispectral image in Plate 3a and a hyperspectral image in
Plate 3b, both courtesy of QuantaLab-
IAS
-
CSIC
, Cordoba, Spain.
Some Color Infrared (
CIR
) cameras are designed with four
spectral channels covering the three R, G, B spectral bands and
the infrared band, i.e., they belong to the category of multispec-
tral systems. Plate 4 displays four strips built with a
CIR
system.
The left and right strips represent
RGB
images; the left central
strip represents a Digital Surface Model (
DSM
) with the associ-
ated color-bar representing surface heights (in meters); the Right
central strip represents the
RGB
plus the IR channel. These im-
ages are courtesy of QuantaLab-
IAS
-
CSIC
, Cordoba, Spain.
Radar/SAR
Synthetic Aperture Radar (
SAR
) systems have been installed
onboard
UAVs
with successful results. Rosen
et al.
(2006)
proposed the design of a polarimetric
SAR
system with a range
bandwidth of 80 MHz for
UAVs
. Wang
at al
. (2009) developed
an operative system based on a technology that combines
millimeter-wave frequency-modulated continuous-wave and
SAR
. The transmission power required is feasible in
UAVs
. This
system appears to be an active image sensor, which can be
used in remote sensing applications, similar to
SAR
onboard
satellites, where land cover for texture analysis is one of
them. Colomina and Molina (2014) provided a list of rep-
resentative
SAR
systems. Zaugg and Long (2008), Xing
et al.
(2009) and later Zhang
et al.
(2012) developed robust motion
compensation approaches for
UAV
SAR
imagery, as appropriate
for highly precise imaging for
UAV
SAR
, which are also valid
for platforms equipped with only a low-accuracy inertial
navigation system.
Ouchi (2013) provided specifications of
SAR
systems
onboard
UAVs
operating at X (7 - 12.5
GHz
) and Ku (12 to 18
GHz
) bands. Koo
et al.
(2012) proposed a
SAR
-based system
operating at the C-Band (ranging from 3.7 to 4.2
GHz
and 5.9
to 6.4
GHz
) and single VV polarization. It was designed for
monitoring soil resources, crops, and trees in agriculture and
forestry. Wang
et al.
(2009), Saldaña and Martinez (2007), or
González-Partida
et al.
(2009) designed and developed
SAR
systems working at the millimeter-wave band with the aim of
transmitting large bandwidth, i.e., high spatial resolution. Al-
though this band could represent a problem caused by motion
errors, that could be larger than the
UAV
operation resolution,
they can be compensated with algorithms such as the Range
Migration Algorithm or using an
IMU
to align target responses
(González-Partida
et al.
, 2009). Two
SAR
-based instruments,
operating in the C- and X-bands, were described in Aguasca
et al.
(2013). They are based on dual receiving channels with
the ability to work in interferometric and polarimetric modes
and equipped with a motion compensation unit to avoid also
motion errors.
Nouvel
et al.
(2007 and 2009) developed a low-cost radar
system to enable avoidance of shading effects produced by
SAR
systems in mountains or urban areas with high density of trees.
With the aim of minimizing the above effects, Weiss
et al.
(2007)
proposed the 3
D ARTINO
(Airborne Radar for Three-dimensional
Imaging and Nadir Observation) imaging radar system.
A radiometer operating in the L-band at 1.4
GHz
was de-
signed in Acevo-Herrera
et al.
(2010) and installed on an
UAV
,
together with a
GPS
and an
IMU
.
The
SARVANT
platform is a fixed-wing aerial vehicle with
a six-meter wingspan and a payload weight of 45 kg (Remy
et al
, 2012; Molina
et al.
, 2013).
SARVANT
is equipped with a
dual-band (X and P) interferometric
SAR
, where the P-band
enables the topographic mapping of densely tree-covered
areas, providing terrain profile information. The combination
of X- and P-band data can be used for biomass estimations. It
is also equipped with a double optical system to cover visible
and
NIR
spectrum.
Schulz (2011) and Essen
et al.
(2012) developed two mil-
limeter wave radar, both operating at 94
GHz
, to be integrated
onboard two unmanned helicopters with payloads of 30/35 kg
and maximum weight of 125/85 kg (with fuel), respectively.
The design of radars was based on the Frequency Modulated
Continuous Wave (
FMCW
) principle, to get the highest possible
average transmission power with the best range of performance.
Chemical Sensors
Berman
et al.
(2012) described the design and adaptation of
the Off-Axis Integrated Cavity Output Spectroscopy (Off-Axis
ICOS) for
UAVs
, primarily described in Paul
et al.
(2001) and
Baer
et al.
(2002), with an operating principle based on in-
frared spectroscopy for measuring water vapor (H
2
O), carbon
dioxide (CO
2
), and methane (CH
4
). This sensor fulfills the
payload requirements for the
UAV
provided by
NASA
(Sensor
Integrated Environmental Remote Research Aircraft, SIERRA),
which is midsize with 6.1 m wingspan, 3.6 m long, and 1.4 m
high with a cruising speed of 28 m/s and a maximum alti-
tude of 3,600 m.
SIERRA
can carry a 40 kg payload measuring
40.5 cm × 40.5 cm × 30.5 cm and can provide up to 200 W of
aircraft power.
A biologically inspired electronic nose is deployed and
installed on an
UAV
in Bermúdez i Badia
et al.
(2007) based
288
April 2015
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
