Evaluation of the gully erosion susceptibility by using UAV and hybrid models based on machine learning
Soil and Tillage Research,
Год журнала:
2024,
Номер
244, С. 106218 - 106218
Опубликована: Июль 5, 2024
Язык: Английский
Utilizing Machine Learning Algorithms for the Development of Gully Erosion Susceptibility Maps: Evidence from the Chotanagpur Plateau Region, India
Sustainability,
Год журнала:
2024,
Номер
16(15), С. 6569 - 6569
Опубликована: Июль 31, 2024
Gully
erosion
is
a
serious
environmental
threat,
compromising
soil
health,
damaging
agricultural
lands,
and
destroying
vital
infrastructure.
Pinpointing
regions
prone
to
gully
demands
careful
selection
of
an
appropriate
machine
learning
algorithm.
This
choice
crucial,
as
the
complex
interplay
various
factors
contributing
formation
requires
nuanced
analytical
approach.
To
develop
most
accurate
Erosion
Susceptibility
Map
(GESM)
for
India’s
Raiboni
River
basin,
researchers
harnessed
power
two
cutting-edge
algorithm:
Extreme
Gradient
Boosting
(XGBoost)
Random
Forest
(RF).
For
comprehensive
analysis,
this
study
integrated
24
potential
control
factors.
We
meticulously
investigated
dataset
200
samples,
ensuring
even
balance
between
non-gullied
gullied
locations.
assess
multicollinearity
among
variables,
we
employed
techniques:
Information
Gain
Ratio
(IGR)
test
Variance
Inflation
Factors
(VIF).
Elevation,
land
use,
river
proximity,
rainfall
influenced
basin’s
GESM.
Rigorous
tests
validated
XGBoost
RF
model
performance.
surpassed
(ROC
86%
vs.
83.1%).
Quantile
classification
yielded
GESM
with
five
levels:
very
high
low.
Our
findings
reveal
that
roughly
12%
basin
area
severely
affected
by
erosion.
These
underscore
critical
need
targeted
interventions
in
these
highly
susceptible
areas.
Furthermore,
our
analysis
characteristics
unveiled
predominance
V-shaped
gullies,
likely
active
developmental
stage,
supported
average
Shape
Index
(SI)
value
0.26
mean
Erosivness
(EI)
0.33.
research
demonstrates
pinpoint
areas
By
providing
valuable
insights,
policymakers
can
make
informed
decisions
regarding
sustainable
management
practices.
Язык: Английский
Waterlogged Area Identification Models Based on Object-Oriented Image Analysis and Deep Learning Methods in Sloping Croplands of Northeast China
Sustainability,
Год журнала:
2024,
Номер
16(10), С. 3917 - 3917
Опубликована: Май 8, 2024
Drainage
difficulties
in
the
waterlogged
areas
of
sloping
cropland
not
only
impede
crop
development
but
also
facilitate
formation
erosion
gullies,
resulting
significant
soil
and
water
loss.
Investigating
distribution
these
is
crucial
for
comprehending
patterns
preserving
black
resource.
In
this
study,
we
built
varied
models
based
on
two
stages
(one
using
deep
learning
methods
other
combining
object-based
image
analysis
(OBIA)
with
methods)
to
identify
high-resolution
remote
sensing
data.
The
results
showed
that
five
original
imagery
achieved
precision
rates
varying
from
54.6%
60.9%.
Among
models,
DeepLabV3+-Xception
model
highest
accuracy,
as
indicated
by
an
F1-score
53.4%.
identified
demonstrated
a
distinction
categories
areas:
zones
risk
areas.
former
had
obvious
borders
fewer
misclassifications,
exceeding
latter
terms
identification
accuracy.
Furthermore,
accuracy
was
significantly
improved
when
combined
object-oriented
analysis.
DeepLabV3+-MobileNetV2
maximum
59%,
which
6%
higher
than
imagery.
Moreover,
advancement
mitigated
issues
related
boundary
blurriness
noise
process.
These
will
provide
scientific
assistance
managing
reducing
impact
places.
Язык: Английский
Machine Learning Models for the Spatial Prediction of Gully Erosion Susceptibility in the Piraí Drainage Basin, Paraíba Do Sul Middle Valley, Southeast Brazil
Land,
Год журнала:
2024,
Номер
13(10), С. 1665 - 1665
Опубликована: Окт. 13, 2024
Soil
erosion
is
a
global
issue—with
gully
recognized
as
one
of
the
most
important
forms
land
degradation.
The
purpose
this
study
to
compare
and
contrast
outcomes
four
machine
learning
models,
Classification
Regression
(CART),
eXtreme
Gradient
Boosting
(XGBoost),
Random
Forest
(RF),
Support
Vector
Machine
(SVM),
used
for
mapping
susceptibility
soil
erosion.
controlling
factors
in
Piraí
Drainage
Basin,
Paraíba
do
Sul
Middle
Valley
were
analysed
by
image
interpretation
Google
Earth
samples
(n
=
159)
modelling
spatial
prediction.
XGBoost
RF
models
achieved
identical
results
area
under
receiver
operating
characteristic
curve
(AUROC
88.50%),
followed
SVM
CART
respectively
86.17%;
AUROC
85.11%).
In
all
analysed,
importance
main
predominated
among
Lineaments,
Land
Use
Cover,
Slope,
Elevation
Rainfall,
highlighting
need
understand
landscape.
model,
considering
smaller
number
false
negatives
prediction,
was
considered
appropriate,
compared
model.
It
noteworthy
that
model
made
it
possible
validate
hypothesis
area,
identifying
9.47%
Basin
susceptible
Furthermore,
replicable
methodologies
are
evidenced
their
rapid
applicability
at
different
scales.
Язык: Английский