Remote Sensing,
Journal Year:
2023,
Volume and Issue:
15(10), P. 2540 - 2540
Published: May 12, 2023
Meeting
current
needs
without
compromising
future
generations’
ability
to
meet
theirs
is
the
only
path
toward
achieving
environmental
sustainability.
As
most
valuable
natural
resource,
soil
faces
global,
regional,
and
local
challenges,
from
quality
degradation
mass
losses
brought
on
by
salinization.
These
issues
affect
agricultural
productivity
ecological
balance,
undermining
sustainability
food
security.
Therefore,
timely
monitoring
accurate
mapping
of
salinization
processes
are
crucial,
especially
in
semi-arid
arid
regions
where
climate
variability
impacts
have
already
reached
alarming
levels.
Salt-affected
has
enormous
potential
thanks
recent
progress
remote
sensing.
This
paper
comprehensively
reviews
sensing
assess
The
review
demonstrates
that
large-scale
salinity
estimation
based
tools
remains
a
significant
challenge,
primarily
due
data
resolution
acquisition
costs.
Fundamental
trade-offs
constrain
practical
applications
between
resolution,
spatial
temporal
coverage,
costs,
high
accuracy
expectations.
article
provides
an
overview
research
work
related
using
By
synthesizing
highlighting
areas
further
investigation
needed,
this
helps
steer
efforts,
insight
for
decision-making
resource
management,
promotes
interdisciplinary
collaboration.
Land,
Journal Year:
2025,
Volume and Issue:
14(1), P. 110 - 110
Published: Jan. 8, 2025
Soil
salinization
affects
agricultural
productivity
and
ecosystem
health
in
Xinjiang,
especially
arid
areas.
The
region’s
complex
topography
limited
data
emphasize
the
pressing
need
for
effective,
large-scale
monitoring
technologies.
Therefore,
1044
soil
samples
were
collected
from
farmland
northern
potential
effectiveness
of
salinity
was
explored
by
combining
environmental
variables
with
Landsat
8
Sentinel-2.
study
applied
four
types
feature
selection
algorithms:
Random
Forest
(RF),
Competitive
Adaptive
Reweighted
Sampling
(CARS),
Uninformative
Variable
Elimination
(UVE),
Successive
Projections
Algorithm
(SPA).
These
are
then
integrated
into
various
machine
learning
models—such
as
Ensemble
Tree
(ETree),
Extreme
Gradient
Boosting
(XGBoost),
LightBoost—as
well
deep
models,
including
Convolutional
Neural
Networks
(CNN),
Residual
(ResNet),
Multilayer
Perceptrons
(MLP),
Kolmogorov–Arnold
(KAN),
modeling.
results
suggest
that
fertilizer
use
plays
a
critical
role
processes.
Notably,
interpretable
model
KAN
achieved
an
accuracy
0.75
correctly
classifying
degree
salinity.
This
highlights
integrating
multi-source
remote
sensing
technologies,
offering
pathway
to
monitoring,
thereby
providing
valuable
support
management.
International Journal of Applied Earth Observation and Geoinformation,
Journal Year:
2022,
Volume and Issue:
112, P. 102969 - 102969
Published: Aug. 1, 2022
Soil
salinization
has
hampered
the
achievement
of
sustainable
development
goals
(SDGs)
in
many
countries
worldwide.
Several
have
recently
launched
hyperspectral
remote
sensing
satellites,
opening
new
avenues
for
accurate
soil-salinity
monitoring.
Among
them,
Gaofen-5
(GF-5)
from
China
a
high
comprehensive
performance,
including
spectral
resolution
5
nm,
330
bands,
and
signal-to-noise
ratio
700.
However,
potential
GF-5
estimating
soil
salinity
is
not
well
understood.
In
this
study,
we
proposed
strategy
that
includes
bootstrap
methods,
fractional
order
derivative
(FOD)
techniques
decision-level
fusion
models
to
exploit
diagnostic
information
reduce
estimation
uncertainty
Ebinur
Lake
oasis
northwestern
China.
The
results
showed
data
were
suitable
assessing
salinity.
FOD
technique
enhanced
correlation
between
spectra,
identified
more
improved
accuracy
estimation,
reduced
model
uncertainty.
low-order
outperformed
high-order
FOD.
spectra
processed
by
0.9
most
correlated
with
(r
=
−0.76).
driven
0.8
produced
optimal
estimated
(R2
0.95,
root
mean
square
error
(RMSE)
3.20
dS
m−1
performance
interquartile
distance
(RPIQ)
5.96).
had
less
than
based
on
original
integer-order
(first-
second-
derivatives)
spectra.
This
study
provides
reference
using
framework
low
accuracy.
great
environmental
problems
facilitating
further
SDGs.
Ecological Indicators,
Journal Year:
2022,
Volume and Issue:
143, P. 109330 - 109330
Published: Aug. 23, 2022
Heavy
metal
pollution
poses
a
huge
challenge
to
the
soil
environment.
With
increasing
level,
traditional
monitoring
methods
cannot
quickly
obtain
information
on
large-area
pollution.
Therefore,
large-scale
mapping
method
with
high
precision
is
urgently
needed
effectively
control
heavy
This
study
explored
for
concentrations
through
hyperspectral
images.
On
this
basis,
new
Stacked
AdaBoost
ensemble
learning
algorithm
was
constructed
construct
inversion
model
of
contents.
The
characteristic
spectral
bands
metals
were
extracted
as
input
variables
using
Pearson's
correlation
coefficient
and
successive
projections
algorithm.
three
sets
content
data,
prediction
accuracy
outcomes
various
machine
compared.
Furthermore,
potential
sources
in
area
analyzed
based
Moran's
index.
results
showed
that
relatively
stable
higher
than
models.
For
Cr,
Cu,
As,
determination
coefficients
(R2)
verification
set
0.66,
0.61,
0.74,
respectively.
Afterward,
used
map
concentration
over
area.
suggested
conditions
soils
Ganhetan
caused
by
nature
human
activities.
As
agricultural
most
serious,
an
exceedance
rate
38.66%.
Industrial
areas
In
summary,
provides
detailed
reliable
data
ecological
protection
industrial
control,
allowing
effective
management
sources.
Geoderma,
Journal Year:
2023,
Volume and Issue:
430, P. 116321 - 116321
Published: Jan. 4, 2023
Soil
salinization
is
a
major
environmental
risk
caused
by
natural
or
human
activities
especially
in
arid
and
semi-arid
regions.
Machine
learning
for
rapidly
monitoring
large-scale
spatial
soil
becomes
possible.
However,
machine
often
needs
large
training
samples
obtaining
extensive
information
field
investigation
laborious
difficult.
In
practice,
the
sampling
datasets
are
sparse
non-normally
distributed.
The
intricacy
of
features
extracted
from
remote
sensing
images
increases
model
complexity
leads
to
degradation
prediction
performance.
To
solve
this
problem,
an
integrative
framework
proposed
predict
salt
content
(SSC)
based
on
light
gradient
boosting
(LGBM).
model,
we
first
introduce
data
augmentation
method
(Mixup)
improve
sample
diversity
alleviate
overfitting
sparsity
samples.
generalization
robustness
different
heterogeneity
salinization,
Mixup-LGBM
adaptively
jointly
optimized
combining
hyperparameters
feature
selection
Bayesian
optimization
framework.
Furthermore,
interpretability
improved
using
shapley
additive
explanations
(SHAP)
value
combination
confidence
synthetic
through
visualization
importance
assessment.
addition,
cases
simulated
test
Case
I,
raw
sample-sparsity
algorithm
has
higher
accuracy
than
other
unused
models.
Ⅱ,
extreme
still
achieves
satisfactory
results
while
models
can’t
learn
any
effective
after
multiple
iterations.
experimental
reveal
that
can
automatically
find
representative
heterogeneous
environments
strong
adaptability
study
areas.
This
finding
indicates
digital
elevation
(DEM)
high
influence
SSC
both
Besides
DEM,
Manasi
River
Basin
more
sensitive
activities,
Werigan–Kuqa
Delta
Oasis
factors.
suitable
predicting
scenarios
ensuring
accuracy.
considerable
potential
dealing
with
complex
regression
tasks.
