PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
April 2016
283
The Attenuation of Retroreflective
Signatures on Surface Soils
Robyn A. Barbato, Karen L. Foley, Adam LeWinter, David Finnegan, Sergey Vecherin, John E. Anderson,
Kenneth Yamamoto, Christian Borden, Ethan Fahy, Nathan Calandra, and Charles M. Reynolds
Abstract
Soil parameters such as water potential, temperature, organic
matter (
OM
), and particle size distribution influence biologi-
cal activity and collectively define the state of soils, yet these
properties are typically described through time-intensive,
ground-based sampling efforts. To improve our understanding
of soils through stand-off sensing techniques, Light Detection
and Ranging was used to monitor the signature of retrore-
flective beads embedded in polymeric agents on four soils.
Our goal was to generate probability density functions (
PDFs
)
for stochastic predictions of the persistence of this signa-
ture through time. Our findings showed that the
PDFs
of the
reflected signal between target and background soils became
nearly indistinguishable after five months and that
OM
,
nitrogen content, cation exchange capacity, and pH related
to signature decline. This approach, while developed using
polymer-bound retroreflectors, will serve as a framework
where a signature-emitter is left in or on soil and differentially
influenced by terrain, weather, and soil processes.
Introduction
Soils are complex systems, exhibiting temporal as well as spa-
tial heterogeneity. There is a critical need to better understand
and predict soil processes subjected to local weather patterns
over large expanses of land and varying terrain. With the ad-
vent of advanced biological tools such as improved sequencing
capabilities and
in situ
respiration techniques providing an im-
mense amount of data, we have a greater understanding of gen-
eral trends in soil activity. The limitation to these techniques
is that they typically require time-intensive ground-based
sampling efforts, often tied to follow-on laboratory analyses.
From a different perspective, modeling may offer predic-
tive capabilities for the complex coupled processes in soils.
Models increasingly use stochastic approaches, such as Monte
Carlo techniques, to try to quantify the effects of uncertain-
ty and variability. Stochastic models benefit from improved
descriptions of the probability density function (
PDF
) of each
process being modeled. Although Gaussian distributions can
be used as default
PDFs
to describe the rates of the processes
involved, not all biochemical processes follow Gaussian dis-
tributions, and the distribution for the rates of a process may
change with the conditions of the system.
Typically, we can generate data that are “deep” in soil
information, but only generate limited sample numbers, or we
can have large sets of data but significantly less information
in each sample. Though the resolution of soil biochemical
processes has improved, it is accompanied by limitations in
generating large sample numbers. However, large sample sets,
often collected by remote sensing techniques, are fundamen-
tal to fully describe
PDFs
used in stochastic predictions. In an
effort to link these two together, we describe an initial study
to relate the rates and patterns of LiDAR signature decay to
soil properties that typically are related to soil biochemical
rates. We investigated coupling LiDAR with emerging soil
analyses to address the limitations of each of the competing
approaches described above.
Lidar sensors collect a dense sampling grid over a large
area of interest in a relatively short period of time by mea-
suring the time of flight of a laser pulse and the azimuth
and zenith angles to obtain x, y, and z coordinates (Eitel
et
al
., 2010). Once the brightness factor was first calibrated by
Kaasalainen
et al
. (2008), reflectance could be measured by
LiDAR sensors. In fact, reflectance measurements over a three
dimensional map can be captured at sub-centimeter scales
using terrestrial LiDAR scanning (
TLS
). Lasers for terrestrial
applications generally have wavelengths that range from 532
to 1550 nm, and within the infrared portion of this range, veg-
etation reflectance is high (Lefsky
et al
., 2002). Traditionally,
TLS
has been used to obtain high resolution terrain data to
identify biomass in vegetated environments (Eitel
et al
., 2010;
H
ӧ
fle, 2014; Kaasalainen
et al
., 2014), reveal rock outcropping
in mountainous regions (Burton
et al
., 2011; Buckley
et al
.,
2013), determine soil surface changes (Sankey
et al
., 2012),
and survey roads and buildings for urban planning (Jaselskis
et al
., 2005; Pu and Vosselman, 2009).
In our study, we exploited
TLS
in a novel way by character-
izing a signature through time and relating its decline to soil
processes that impact activity. This technique could be applied
to detect changes in a landscape, from natural and anthro-
pogenic occurrences such as erosion and intruder detection,
respectively. The difficulty in obtaining accurate soil activity
measurements also results in limitations to soil modeling ap-
proaches, particularly stochastic techniques that rely on mean-
ingful descriptions of the variability in the input parameters to
provide uncertainty estimates in the output. Given these gaps
in modeling capabilities, this work was specifically aimed to
maximize the sample set by using
TLS
as a non-invasive tool to
characterize signature decline on surface soils through time. In
our study, microscopic retroreflective beads were embedded on
surface soils using a dust abatement polymer. Because retrore-
flectors return light back to the source with minimum scattering
and are distributed in small quantities, they can be separated
Adam LeWinter, David Finnegan, Sergey Vecherin, Kenneth
Yamamoto, Charles M. Reynolds, Robyn Barbato, Karen
Foley, and David Finnegan are with the US Army ERDC-Cold
Regions Research and Engineering Laboratory, 72 Lyme Road,
Hanover, NH 03755 (
).
John E. Anderson is the US Army ERDC-Geospatial Research
Laboratory, 1000 W. Cary Street (VCU Academic Campus)
Richmond, VA 23284.
Christian Borden, Ethan Fahy, and Nathan Calandra are with
the Atmospheric and Environmental Research, 131 Hartwell
Avenue, Lexington, MA 02421.
Photogrammetric Engineering & Remote Sensing
Vol. 82, No. 4, April 2016, pp. 283–289.
0099-1112/16/283–289
© 2016 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.82.4.283
