An Evaluation of Reflectance Calibration
Methods for UAV Spectral Imagery
Jarrod Edwards, John Anderson, William Shuart, and Jason Woolard
Abstract
Spectral imagery using micro-unmanned aerial vehicles is
rapidly advancing. This study compared reflectance calibra-
tion methods for imagery acquired using the Parrot Sequoia
imager, a commercial multispectral sensor package. For the
study, two orthomosaics were calibrated using 1) a manufac-
turer-suggested
AIRINOV
standar
ware and 2) the Empirical Line
ground radiometric data on spe
scenes were analyzed for target
radiometric survey. Regression analysis demonstrated more
favorable target correlation for the
ELC
imagery than the
AIRI-
NOV
-calibrated imagery, with Root Mean Square Error (
RMSE
)
analysis supporting these results. Finally, classification maps
were produced between the data sets. Error analysis resulted
in an overall accuracy of 24% for the
AIRINOV
map compared
to
ELC
-based truth data, with a considerable number of pixels
associated with brighter targets unclassified. These results
demonstrate the need for standardized calibration procedures
in the spectral correction of small-format remote sensor data.
Key words: micro-
UAS
,
UAV
, drone, remote sensing, spectral
reflectance, empirical line calibration, classification, ground
radiometry, error analysis, multispectral.
Introduction
The data produced from air and satellite-based remote sen-
sor systems has dominated the geospatial sciences for many
decades with a rich variety of spatial and spectral imagery
formats. These data have contributed significantly to the
social, natural and physical science disciplines with products
of varied quality and utility for a variety of mapping applica-
tions (Haboudane
et al.
2004; Mumby
et al.
1998; Cowley
et
al.
2018). Many of these data are also largely verifiable and
collected using standardized, well-documented correction
procedures that an end-user can make reference to when
exploiting and certifying the data or resulting products (Price
1987; Cocks
et al.
1998). For example, Thematic Mapper
is available in various radiometrically corrected configura-
tions and post-processed product levels that are rigorously
researched and provide the user an idea of the quality of the
data with relation to spectral fidelity (Thome
et al.
1993;
Chander
et al.
2009; N. Mishra
et al.
2016; Chander and
Markham 2003; Chander
et al.
2007). In addition, airborne
multi- and hyperspectral data are provided in formats that
typically report calibration methods that include atmospheric
correction routines and reflectance calibration procedures
(Richter and Schläpfer 2002; Eismann 2012; Gao
et al.
2009).
Until recently, data integrity has been discussed and present-
ed as a verifiable and traceable part of many of the systems
remote sensing has depended upon (Schaepman-Strub 2006).
de has seen the explosive development
s paired with unmanned aerial sys-
at have offered greater flexibility and
capabilities in spatial, spectral, and
temporal resolution (Elias 2012; Colomina and Molina 2014).
These imaging systems can include (non-metric) compact to
sub-compact cameras of various quality that produce frame
imagery collected using a combination of on-board global
positioning system (
GPS
) and inertial measurement system
triggers that are post processed using analytical photogram-
metry. The benefits and flexibility
UAS
offers includes size,
temporal exploitation, and an ever-widening availability
of sensor payloads. But while the size and weight of many
systems are convenient for transport and deployment, it does
not typically translate into a stable sensor platform in the air.
In fact, many systems are highly unstable in flight serving to
introduce post-processed spatial and radiometric errors as-
sociated with changing viewing geometry and photogrammet-
ric exterior orientation issues. These changes often include
topography and sun angle issues manifested as bidirectional
reflectance distribution function (
BRDF
) variations in the indi-
vidual images comprising a final image ortho-mosaic (Burkart
et al.
2015; Stark
et al.
2016). Both the potential platform sta-
bility issues and a shortage of research on the spectral fidelity
of these new compact sensors beg for validation standards for
drone-based electro-optic (
EO
) data similar to those described
by Justice
et al.
(2010) for satellite sensor products.
There are a number of
UAS
-based multispectral sensors that
are currently available. Tetracam offers a 6-band
MCA6
mul-
tispectral imager covering the 450–1000 nm spectral region
with a 38°-by-31° field-of-view. The MicaSense RedEdge-M
sensor is a 5-band system (band centers at 475, 560, 668, 717,
and 840 nm) that produces 12-bit
RAW
imagery with a 47.2°
horizontal field-of-view (
HFOV
) (Barrows and Bulanon 2017;
Potgeiter
et al.
2017). Comparably, Sentera has developed the
Multispectral Double 4K developed by Parrot (Paris, France)
that is among these recently-released micro (remote) sensors.
Primarily developed as an agricultural sensor, it has a five-
lens architecture—four narrowband sensors (band centers at
550, with five channels (band centers at 446, 548, 650, 720,
840 nm) that utilizes a 12.3
MP
backside illumination (
BSI
)
complementary metal oxide semiconductor (
CMOS
) with a 60°
HFOV
. The Slantrange 3p imager is another four-channel sys-
tem with configurable band centers between 410 and 950 nm.
Jarrod Edwards and John Anderson are with the USACE–
ERDC Geospatial Research Laboratory, Corbin Field
Station, 15319 Magnetic Lane, Woodford, Va. 22580
(
). (John.Anderson@usace.
army.mil).
William Shuart is with the Center for Environmental Studies,
Virginia Commonwealth University, 1000 W. Cary St,
Richmond Va. 22580
).
Jason Woolard is with the NOAA National Ocean Service,
National Geodetic Survey, 15351 Office Dr, Woodford Va.
22580 (
).
Photogrammetric Engineering & Remote Sensing
Vol. 85, No. 3, March 2019, pp. 221–230.
0099-1112/18/221–230
© 2019 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.85.3.221
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
March 2019
221
