Buildings,
Год журнала:
2024,
Номер
14(9), С. 2616 - 2616
Опубликована: Авг. 23, 2024
The
construction
field
currently
suffers
from
low
productivity,
a
lack
of
expertise
among
practitioners,
weak
innovation,
and
predictability.
digital
twin,
an
advanced
technology,
empowers
the
sector
to
advance
towards
intelligent
transformation.
It
ultimately
aims
for
highly
accurate
simulation
achieve
comprehensive
optimization
all
phases
project.
Currently,
process
twin
applications
is
facing
challenges
such
as
poor
data
quality,
inability
harmonize
types
that
are
difficult
integrate,
insufficient
security.
Further
research
on
application
twins
in
domain
still
needed
accelerate
development
promote
their
practical
application.
This
paper
analyzes
commonly
used
architectures
literature
summarizes
technologies
implement
architectures,
including
artificial
intelligence,
machine
learning,
mining,
cyber–physical
systems,
internet
things,
virtual
reality,
augmented
reality
applications,
considers
advantages
limitations.
focus
this
centered
entire
lifecycle
project,
which
includes
design,
construction,
operation,
maintenance,
demolition
restoration
phases.
Digital
mainly
moving
integration
information,
model
automation,
system
control,
security
privacy.
present
management
challenges,
privacy
protection,
technical
manpower
development,
transformation
needs.
Future
should
address
these
by
improving
developing
robust
methodologies,
strengthening
measures.
Forests,
Год журнала:
2024,
Номер
15(8), С. 1374 - 1374
Опубликована: Авг. 6, 2024
Trees’
structural
defects
are
responsible
for
the
reduction
in
forest
product
quality
and
accident
of
tree
collapse
under
extreme
environmental
conditions.
Although
manual
view
inspection
assessing
health
condition
is
reliable,
it
inefficient
discriminating,
locating,
quantifying
with
various
features
(i.e.,
crack
hole).
There
a
general
need
investigation
efficient
ways
to
assess
these
enhance
sustainability
trees.
In
this
study,
deep
learning
algorithms
lightweight
You
Only
Look
Once
(YOLO)
encoder-decoder
network
named
DeepLabv3+
combined
unmanned
aerial
vehicle
(UAV)
observations
evaluate
trees’
defects.
Experimentally,
we
found
that
state-of-the-art
detector
YOLOv7-tiny
offers
real-time
50–60
fps)
long-range
sensing
5
m)
but
has
limited
capacity
acquire
patterns
at
millimeter
scale.
To
address
limitation,
further
utilized
cascaded
different
architectures
ResNet18,
ResNet50,
Xception,
MobileNetv2
obtain
actual
morphology
through
close-range
pixel-wise
image
semantic
segmentation.
Moreover,
proposed
hybrid
scheme
YOLOv7-tiny_DeepLabv3+_UAV
assesses
tree’s
defect
size
an
averaged
accuracy
92.62%
(±6%).
Buildings,
Год журнала:
2024,
Номер
14(9), С. 2616 - 2616
Опубликована: Авг. 23, 2024
The
construction
field
currently
suffers
from
low
productivity,
a
lack
of
expertise
among
practitioners,
weak
innovation,
and
predictability.
digital
twin,
an
advanced
technology,
empowers
the
sector
to
advance
towards
intelligent
transformation.
It
ultimately
aims
for
highly
accurate
simulation
achieve
comprehensive
optimization
all
phases
project.
Currently,
process
twin
applications
is
facing
challenges
such
as
poor
data
quality,
inability
harmonize
types
that
are
difficult
integrate,
insufficient
security.
Further
research
on
application
twins
in
domain
still
needed
accelerate
development
promote
their
practical
application.
This
paper
analyzes
commonly
used
architectures
literature
summarizes
technologies
implement
architectures,
including
artificial
intelligence,
machine
learning,
mining,
cyber–physical
systems,
internet
things,
virtual
reality,
augmented
reality
applications,
considers
advantages
limitations.
focus
this
centered
entire
lifecycle
project,
which
includes
design,
construction,
operation,
maintenance,
demolition
restoration
phases.
Digital
mainly
moving
integration
information,
model
automation,
system
control,
security
privacy.
present
management
challenges,
privacy
protection,
technical
manpower
development,
transformation
needs.
Future
should
address
these
by
improving
developing
robust
methodologies,
strengthening
measures.
