three blimps of nylon with helium bottles for inflating with
lengths of 12.4 m to 18.3m and payload between 6 to 15 kg;
and (b) three fixed-wing models, with lengths of 2.0 m to 2.8
m with a payload between 2 to 5 kg.
Whitehead
et al.
(2013) measured surface motion and
elevation changes of an Arctic glacier based on image process-
ing techniques. They used
UAVs
and then a piloted helicopter.
A
DEM
and an ortho-mosaic were generated with accuracies
comparable to those obtained from a fixed-wing
UAV
with
payload of approx. 0.5 kg.
Bathymetry measurements, obtained with very high spatial
resolution imagery sensors onboard
UAVs
, were obtained for
producing
DEMs
and studying the riverbed of the Ain and the
Drôme rivers in France (Lejot
et al.
2007).
Aerial photography was combined with mobile lidar on
the ground for obtaining a seamless
DTM
with the purpose
of studying changes in river channels and their floodplains
(Flener
et al.
2007).
Yun
et al.
(2012) and Kim
et al.
(2013) proposed the use of
a 3G Smart phone onboard a fixed-wing
UAV
, with all devices
previously calibrated, including the camera for photogram-
metric purposes. The system was programmed based on the
Android operating system.
Matsuoka
et al.
(2012) applied photogrammetry-based
techniques for deformation measurements, with sufficient
accuracy, of a large-scale solar power plant. They used im-
ages acquired with a calibrated, non-digital camera onboard a
quad-copter and based on a sufficient number of ground con-
trol points. Control points are relevant features in photogram-
metry for product generation as tested in Chiang
et al.
(2012).
An
UAV
with wingspan of 5 m and payload of 25 kg was used.
Manyoky
et al.
(2011) reported that
UAVs
were tested for
capturing geodata and compared with conventional acqui-
sition methods for cadastral applications. Two different
methods (tachymetry/
GNSS
and an
UAV
) were applied in two
test areas in Switzerland: (a) a mountainous area in Krattigen;
and (b) a suburban area in Campus Science City
ETH
Zurich.
Regarding the
UAV
, an octo-copter with payload of 500 g was
used. It was equipped with a
GNSS
, a barometric height sensor,
a compass, an
IMU
, and a commercial camera weighing 265 g.
Cadastral surveys were carried out by Cunningham
et al.
(2011) in rural areas in Alaska for generating orthomosaics,
from an
UAV
-mounted camera, with the purpose of testing
other sensors (lidar,
SAR
) also for cadastral mapping. Hinsberg
et al.
(2013) used images acquired from an
UAV
for boundary
identification, obtaining sufficient accuracies in the experi-
ments carried out in Austerlitz and the City of Nunspeet,
Netherlands.
DEMs
quality represents a challenge for cadastral
applications. In this regard, Berteška and Ruzgienė (2013)
have analyzed this issue using a fixed-wing
UAV
platform
with wingspan of 1.8 m and a take-off weight around 4 kg,
equipped with an off-the-shelf
CCD
-based camera.
Immerzeel
et al.
(2014) applied stereovision and
SfM
to de-
rive highly detailed ortho-mosaics and
DEMs
by using a fixed-
wing
UAV
with a wingspan of 80 cm and a take-off capacity
of 0.5 kg. It is equipped with a digital camera triggered by the
autopilot system. The goal was the observation of debris-cov-
ered tongue of the Lirung Glacier in Nepal before and after the
melt and monsoon season. Stereovision is also the technique
applied in Stefanik
et al.
(2011) for terrain 3
D
mapping. The
stabilized stereovision system consisted of two gray cameras
with a baseline of 1.5 m plus an
IMU
with total weight of 28 kg
onboard an autonomous helicopter with payload of 94 kg.
DSMs
products were also obtained, with accuracy and
precision, in Tonkin
et al.
(2014) from a large glacial circle at
Cwm Idwal, North Wales, by applying
SfM
to images acquired
with a hexa-copter equipped with a digital
RGB
commer-
cial camera. High-resolution spatial analysis of mountain
landscapes was also carried out in Wundram and Loffler
(2008) based on
DEM
production.
Mosaicking, Ortho- and Geo-rectification
Images acquired with
UAVs
cover relatively small land areas;
hence, automatic mosaicking and rectification are required
for covering larger areas when required. Most applications
described above were based on mosaic and ortho-rectified
images. Here, only specific aspects regarding both topics are
addressed.
Mosaicking and ortho-rectification are products obtained
in real-time from video streams (Zhou, 2009), where the
UAV
platform has 1.53 m (length) × 1.53 m (height), with wingspan
of 2.44 m, weight of 10 kg and payload of 2.3 kg. Different
efficient methods have been proposed for accuracy in the gen-
eration of such products, including
SIFT
for matching (Xing
et al.
, 2010a and 2010b; Xing, 2010; Yang
et al.
, 2013). The
efficiency in the generation of these products with
UAVs
, with
relatively high ease of use, has led to considerable advances
from the point of view of remote sensing in the last few years
(Zhao
et al.
, 2006).
SIFT
is also the method used in Zhang
et
al.
(2011) with overlapped images captured from
UAVs
to pro-
duce photogrammetric products.
Feature matching and
SfM
were used in Turner
et al.
(2012)
for geometric correction and mosaicking. The images were
processed to create three-dimensional point clouds, which are
used to build
DTMs
; then, images are mosaicked. An octo-cop-
ter, with a digital commercial camera, is the
UAV
used with
payload limit of approximately 1 kg.
Turner
et al.
(2014a) applied a direct georeferencing tech-
nique by synchronizing the camera exposure time with the
position of each airframe recorded by a
GPS
. Image processing
techniques were used to eliminate blurry images and images
with excessive overlapping. They compared three different
software methods using an octo-copter with payload capability
of up to 2 kg, equipped with a commercial and stabilized digi-
tal camera. Direct image georeferencing using the
UAVs
georef-
erencing was previously tested in Bláha
et al.
(2011) using and
octo-copter with payload of 500 g and equipped with a camera,
three magnetometers, a barometric altimeter, an
INS
and a
GPS
.
In the context of agriculture, an automatic image-based
UAV
system was developed in Xiang and Tian (2011) to obtain
georeferenced images for mosaicking and ortho-rectification.
The system was a helicopter, a
CMOS
multispectral camera
ranging from 520 nm to 950 nm, an
IMU
, a differentially-cor-
rected
GPS
, a single board computer (
SBC
), a flight controller, a
pulse-width modulation switch, a wireless router, and a video
transmitter.
Mayr (2011) reported on the applicability of a 1.1 kg fixed
wing
UAV
in photogrammetry for building ortho-mosaics and
DSMs
through several examples, and Gülch, (2012) for photo-
grammetric measurements using different software packages.
Regarding the orthorectification, Mesas-Carrascosa
et al.
(2014) have studied the positional quality of orthophotos
obtained with an off-the-shelf commercial digital camera
onboard a quad-copter with maximum payload of 1.2 kg.
Aerial triangulation,
DSM
generation, and mosaicking were the
processing techniques applied.
Because of the large amount of data provided by
UAVs
equipment, Stødle
et al.
(2014) proposed a system for 3
D
data
visualization based on raster maps and topography with high
performance.
Atmospheric: Observation, Air Analysis, and Pollution
New challenges in atmospheric observations are now steadily
growing within the area of
UAVs
, where some vehicles and
sensors have been specifically designed for such purpose.
Air concentrations of gases, aerosols, ozone concentrations,
temperature, humidity, pressure, wind fields are the most
important data collected.
306
April 2015
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
