Using YOLO Object Detection to Identify Hare and Roe Deer in Thermal Aerial Video Footage—Possible Future Applications in Real-Time Automatic Drone Surveillance and Wildlife Monitoring
Published: Nov. 27, 2023
Wildlife
monitoring
can
be
time-consuming
and
expensive,
but
the
fast-developing
technologies
of
uncrewed
aerial
vehicles,
sensors,
machine
learning
pave
way
for
automated
monitoring.
In
this
study
we
trained
YOLOv5
neural
networks
to
detect
Points
Interest,
hare
(Lepus
europaeus),
roe
deer
(Capreolus
capreolus)
in
thermal
footage
proposed
a
method
manually
assess
parameter
mean
average
precision
(mAP),
compared
number
actual
false
positive
negative
detections
subsample.
This
showed
that
mAP
close
1
model
does
not
necessarily
perfect
detection
provided
gain
insights
into
parameters
affecting
models'
precision.
Furthermore,
basic,
conceptual
algorithm
implementing
real-time
object
aircraft
systems
equipped
with
high
zoom
capabilities,
laser
rangefinder.
Real-time
is
becoming
an
invaluable
complementary
tool
cryptic
nocturnal
animals
use
sensors.
Language: Английский
Wildlife monitoring with drones: A survey of end users
Wildlife Society Bulletin,
Journal Year:
2024,
Volume and Issue:
48(3)
Published: June 24, 2024
Abstract
Rapid
advancements
in
technology
often
yield
research
inquiry
into
novel
applications
and
drone
(i.e.,
unoccupied
aircraft
systems
or
UAS)
wildlife
management
are
no
exception.
We
questioned
the
time
lag
between
drone‐related
end‐user
assessments.
implemented
an
online,
cross‐sectional
survey
of
professionals
to
better
understand
current
use
benefits
concerns,
complemented
by
a
review
contemporary
peer‐reviewed
gray
literature.
found
little
disparity
scientific
experiences
similar
trends
among
concerns
published
literature
results).
Exploring
new
computer
vision)
refining
original
evaluating
animal
behavior
responses
during
monitoring)
were
strong
pilots
relatively
minimal
experience
(1–5
years).
Advancements
changes
legislation
will
continue
offer
challenges.
Language: Английский