Deep learning-enabled exploration of global spectral features for photosynthetic capacity estimation
Frontiers in Plant Science,
Год журнала:
2025,
Номер
15
Опубликована: Янв. 13, 2025
Spectral
analysis
is
a
widely
used
method
for
monitoring
photosynthetic
capacity.
However,
vegetation
indices-based
linear
regression
exhibits
insufficient
utilization
of
spectral
information,
while
full
spectra-based
traditional
machine
learning
has
limited
representational
capacity
(partial
least
squares
regression)
or
uninterpretable
(convolution).
In
this
study,
we
proposed
deep
model
with
enhanced
interpretability
based
on
attention
and
indices
calculation
global
feature
mining
to
accurately
estimate
We
explored
the
ability
uncover
optimal
form
illustrated
its
advantage
over
methods.
Furthermore,
verified
that
power
compression
was
an
effective
processing.
Our
results
demonstrated
new
outperformed
models,
increase
in
coefficient
determination
(R
2
)
0.01-0.43
decrease
root
mean
square
error
(RMSE)
1.58-12.48
μmol
m
-2
s
-1
.
The
best
performance
our
R
0.86
0.81
maximum
carboxylation
rate
(
V
cmax
electron
transport
J
max
),
respectively.
photosynthesis-sensitive
bands
identified
by
were
predominantly
visible
range.
most
sensitive
discovered
Reflectancenear−infrared+Reflectancegreen/blueReflectancenear−infrared×Reflectancered
provides
framework
interpreting
information
estimating
Язык: Английский
Water content estimation of conifer needles using leaf-level hyperspectral data
Frontiers in Plant Science,
Год журнала:
2024,
Номер
15
Опубликована: Сен. 6, 2024
Water
is
a
crucial
component
for
plant
growth
and
survival.
Accurately
estimating
simulating
water
content
can
help
us
promptly
monitor
the
physiological
status
stress
response
of
vegetation.
In
this
study,
we
constructed
loss
curves
three
types
conifers
with
morphologically
different
needles,
then
evaluated
applicability
12
commonly
used
indices,
finally
explored
leaf
estimation
from
hyperspectral
data
needles
various
morphology.
The
results
showed
that
rate
Olgan
larch
approximately
8
times
higher
than
Chinese
fir
pine
21
Korean
pine.
reflectance
changes
were
most
significant
in
near
infrared
region
(NIR,
780-1300
nm)
short-wave
(SWIR,
1300-2500
nm).
sensitive
bands
conifer
mainly
concentrated
SWIR
region.
indices
suitable
single
type
needles.
partial
least
squares
regression
(PLSR)
model
effective
all
morphologies
demonstrating
PLSR
promising
tool
content.
Язык: Английский
An atmospheric correction method for Himawari-8 imagery based on a multi-layer stacking algorithm
Ecological Informatics,
Год журнала:
2025,
Номер
unknown, С. 103001 - 103001
Опубликована: Янв. 1, 2025
Язык: Английский
A hybrid framework for estimating photovoltaic dust content based on UAV hyperspectral images
International Journal of Applied Earth Observation and Geoinformation,
Год журнала:
2025,
Номер
139, С. 104500 - 104500
Опубликована: Март 27, 2025
Язык: Английский
Nitrogen monitoring and inversion algorithms of fruit trees based on spectral remote sensing: a deep review
Frontiers in Plant Science,
Год журнала:
2024,
Номер
15
Опубликована: Ноя. 22, 2024
Nitrogen,
as
one
of
the
important
elements
affecting
growth
and
development
fruit
trees,
leads
to
slowed
protein
synthesis
reduced
photosynthesis,
resulting
in
yellowing
leaves,
poor
tree
growth,
decreased
yield
under
nitrogen-deficient
conditions.
In
order
minimize
losses
maximize
yield,
there
is
often
an
occurrence
excessive
fertilization,
soil
structure
degradation,
water
pollution.
Therefore,
accurate
real-time
monitoring
nitrogen
content
trees
has
become
fundamental
prerequisite
for
precision
management
orchards.
Furthermore,
orchard
crucial
enhancing
quality
by
maintaining
optimal
conditions
necessary
trees.
Moreover,
it
plays
a
vital
role
safeguarding
ecological
environment
mitigating
overuse
fertilizers
pesticides.
With
continuous
application
spectral
remote
sensing
technology
agricultural
land
management,
this
can
provide
effective
method
content.
Based
on
review
relevant
literature,
paper
summarizes
research
framework
inversion
which
provides
help
further
research.
Firstly,
based
different
platforms,
was
discussed,
acquisition
Secondly,
index
parameters
that
reflect
are
summarized,
practical
guidance
monitoring.
Additionally,
regression
algorithms
situations
data
were
introduced.
conclusion,
response
current
issues
technological
limitations,
future
should
focus
studying
characteristics
during
phenological
periods,
integrating
multi-type
information,
thereby
improving
universality
model
Язык: Английский