Journal of Applied Remote Sensing,
Journal Year:
2024,
Volume and Issue:
18(04)
Published: Nov. 19, 2024
The
semantic
imbalance
of
class
boundary
areas
is
a
key
factor
in
decreasing
the
classification
accuracy
remote
sensing
land
cover
algorithm.
We
propose
multi-source
image
segmentation
network
based
on
multi-modal
collaboration
and
boundary-guided
fusion
(BGF).
BGF
module
uses
information
as
restriction
condition,
embeds
alignment
strategies
into
encoder,
enhances
deep
features
each
mode.
On
this
basis,
guidance
strategy
used
to
assign
different
weights
internal
area
category
guide
feature
fusion.
Furthermore,
reduce
impact
heterogeneity
fusion,
cross-modal
collaborative
constructed
associate
complementary
between
fully
explore
relationship
images
from
both
spatial
channel
domains.
comparative
experiments
were
conducted
with
representative
algorithms
WHU-OPT-SAR
data
set.
experimental
results
show
that
proposed
method
has
increased
mean
intersection
over
union
overall
indicators
by
3.3%
2.2%,
respectively,
compared
MCANet,
especially
road
merger
ratio
10.0%
MCANet.
proved
effectiveness
model.
The Science of The Total Environment,
Journal Year:
2025,
Volume and Issue:
978, P. 179428 - 179428
Published: April 16, 2025
This
study
explores
methane
emission
trends
across
Greece,
Armenia,
and
Rostov
Oblast
region
of
Russia
from
2004
to
2023.
Our
analyses,
based
on
remote
sensing
advanced
statistical
techniques,
showed
a
1.3-1.8
°C
increase
in
mean
annual
temperature
over
this
20-year
period
all
these
three
regions,
with
the
highest
lowest
rates
warming
Armenia
(0.104
°C)
(0.052
°C),
respectively.
Mean
concentrations
increased
distinctly
regions
period.
Greece
trend
correlations
between
emissions
temperatures,
including
seasonal
highlighting
substantial
role
climate
change
trends.
The
on-ground
observations
revealed
intricate
connections
reduced
precipitations,
farming
practices,
waste
disposal
methods,
naturally
occurring
Greece.
In
contrast,
exhibited
weak
emissions,
its
farming,
management,
energy
manufacturing
sectors
playing
significant
determining
quantities.
demonstrated
weaker
association
temperatures
than
being
primarily
shaped
by
agricultural
activities
natural
discharges
wetlands.
forecast
models
predicted
further
rise
7-year
(2024-2030),
elevation
rate
estimated
for
Russia.
emphasizes
need
tailored
mitigation
strategies
address
effectively,
considering
region-specific
factors.
Advanced
monitoring
technologies
provide
crucial
insights
into
assessment
management
diverse
geomorphological
regions.
Open Geosciences,
Journal Year:
2024,
Volume and Issue:
16(1)
Published: Jan. 1, 2024
Abstract
Normalized
difference
vegetation
index
(NDVI)
and
land
surface
temperature
(LST)
are
important
indicators
of
ecological
changes,
their
spatial
temporal
variations
coupling
can
provide
a
theoretical
basis
for
the
sustainable
development
environment.
Based
on
MOD13A1
MOD11A2
datasets,
distribution
characteristics
NDVI
LST
from
2000
to
2020
were
analyzed,
trend
change
slope
method
model
used
calculate
significant
changes.
Finally,
was
degree
between
LST.
The
study
shows
that:
(1)
From
2020,
annual
value
Mu
Us
Sandy
Land
0.25
0.43,
showing
stable
upward
overall,
with
an
increase
rate
0.074/(10a).
proportion
improvement
areas
in
area
is
81.48%.
(2)
There
differences
Land,
overall
decreasing
northwest
southeast
higher
west
than
east.
greatly
affected
by
changes
use
types.
spatiotemporal
variation
different
gradual
warming
global
climate
change.
main
reason
that
human
activities
have
changed
types
increased
local
coverage.
(3)
negative
correlation
R
2
0.5073
passing
significance
test
at
0.01
level.
This
indicates
engineering
policies
effectively
reduce
area,
thereby
achieving
effect
improving
very
high
level,
average
0.895
area.
two
mainly
exhibit
state
mutual
antagonism
space,
reflecting
importance
green
regulating
regional
result
joint
influence
change,
dominated
2020.
Sensors,
Journal Year:
2024,
Volume and Issue:
24(20), P. 6713 - 6713
Published: Oct. 18, 2024
Human
activity
recognition
is
known
as
the
backbone
of
development
interactive
systems,
such
computer
games.
This
process
usually
performed
by
either
vision-based
or
depth
sensors.
So
far,
various
solutions
have
been
developed
for
this
purpose;
however,
all
challenges
not
completely
resolved.
In
paper,
a
solution
based
on
pattern
has
labeling
and
scoring
physical
exercises
in
front
Kinect
sensor.
Extracting
features
from
human
skeletal
joints
then
generating
relative
descriptors
among
them
first
step
our
method.
led
to
quantification
meaningful
relationships
between
different
parts
during
exercise
performance.
method,
discriminating
each
motion
are
used
identify
adaptive
kernels
Constrained
Energy
Minimization
method
target
detector
operator.
The
results
indicated
an
accuracy
95.9%
motions.
Scoring
motions
was
second
after
process,
which
geometric
interpolate
numerical
quantities
extracted
descriptor
vectors
transform
into
semantic
scores.
demonstrated
coincided
with
scores
derived
sports
coach
99.5
grade
R
