Journal of Imaging,
Journal Year:
2024,
Volume and Issue:
11(1), P. 5 - 5
Published: Dec. 31, 2024
The
geometric
feature
characterization
of
fruit
trees
plays
a
role
in
effective
management
orchards.
LiDAR
(light
detection
and
ranging)
technology
for
object
enables
the
rapid
precise
evaluation
features.
This
study
aimed
to
quantify
height,
canopy
volume,
tree
spacing,
row
spacing
an
apple
orchard
using
three-dimensional
(3D)
sensor.
A
sensor
was
used
collect
3D
point
cloud
data
from
orchard.
Six
samples
trees,
representing
variety
shapes
sizes,
were
selected
collection
validation.
Commercial
software
python
programming
language
utilized
process
collected
data.
processing
steps
involved
conversion,
radius
outlier
removal,
voxel
grid
downsampling,
denoising
through
filtering
erroneous
points,
segmentation
region
interest
(ROI),
clustering
density-based
spatial
(DBSCAN)
algorithm,
transformation,
removal
ground
points.
Accuracy
assessed
by
comparing
estimated
outputs
with
corresponding
measured
values.
sensor-estimated
heights
3.05
±
0.34
m
3.13
0.33
m,
respectively,
mean
absolute
error
(MAE)
0.08
root
squared
(RMSE)
0.09
linear
coefficient
determination
(r2)
0.98,
confidence
interval
(CI)
−0.14
−0.02
high
concordance
correlation
(CCC)
0.96,
indicating
strong
agreement
accuracy.
volumes
13.76
2.46
m3
14.09
2.10
m3,
MAE
0.57
RMSE
0.61
r2
value
0.97,
CI
−0.92
0.26,
demonstrating
precision.
For
distances
3.04
0.17
3.18
0.24
3.35
3.40
0.05
values
0.12
0.92
0.07
0.94
respectively.
−0.18
0.01,
−0.1,
0.002
Although
minor
differences
observed,
estimates
efficient,
though
specific
measurements
require
further
refinement.
results
are
based
on
limited
dataset
six
values,
providing
initial
insights
into
performance.
However,
larger
would
offer
more
reliable
accuracy
assessment.
small
sample
size
(six
trees)
limits
generalizability
findings
necessitates
caution
interpreting
results.
Future
studies
should
incorporate
broader
diverse
validate
refine
characterization,
enhancing
practices
Remote Sensing,
Journal Year:
2024,
Volume and Issue:
16(23), P. 4370 - 4370
Published: Nov. 22, 2024
This
paper
presents
an
extensive
review
of
techniques
for
plant
feature
extraction
and
segmentation,
addressing
the
growing
need
efficient
phenotyping,
which
is
increasingly
recognized
as
a
critical
application
remote
sensing
in
agriculture.
As
understanding
quantifying
structures
become
essential
advancing
precision
agriculture
crop
management,
this
survey
explores
range
methodologies,
both
traditional
cutting-edge,
extracting
features
from
images
point
cloud
data,
well
segmenting
organs.
The
importance
accurate
phenotyping
underscored,
given
its
role
improving
monitoring,
yield
prediction,
stress
detection.
highlights
challenges
posed
by
complex
morphologies
data
noise,
evaluating
performance
various
emphasizing
their
strengths
limitations.
insights
offer
valuable
guidance
researchers
practitioners
fields
science
experimental
section
focuses
on
three
key
tasks:
3D
generation,
2D
image-based
extraction,
shape
classification,
segmentation.
Comparative
results
are
presented
using
collected
several
publicly
available
datasets,
along
with
insightful
observations
inspiring
directions
future
research.
Research Square (Research Square),
Journal Year:
2024,
Volume and Issue:
unknown
Published: Dec. 17, 2024
Abstract
Accurate
crop
row
detection
is
an
important
foundation
for
agricultural
machinery
to
realize
autonomous
operation.
In
this
paper,
a
real-time
soybean-corn
method
based
on
GD-YOLOv10n-seg
with
PCA
fitting
proposed.
Firstly,
the
dataset
of
was
established,
and
image
labeled
by
line
label.
Then,
improved
model
constructed
integrating
GhostModule
DynamicConv
into
YOLOv10n-segmentation
model.
The
experimental
results
show
that
performs
better
in
MPA
MiOU,
size
reduced
18.3%.
center
segmentation
fitted
PCA,
accuracy
reaches
95.08%,
angle
deviation
1.75°,
overall
processing
speed
57.32FPS.
This
study
can
provide
efficient
reliable
solution
navigation
operations
such
as
weeding
pesticide
application
under
compound
planting
mode.
Journal of Imaging,
Journal Year:
2024,
Volume and Issue:
11(1), P. 5 - 5
Published: Dec. 31, 2024
The
geometric
feature
characterization
of
fruit
trees
plays
a
role
in
effective
management
orchards.
LiDAR
(light
detection
and
ranging)
technology
for
object
enables
the
rapid
precise
evaluation
features.
This
study
aimed
to
quantify
height,
canopy
volume,
tree
spacing,
row
spacing
an
apple
orchard
using
three-dimensional
(3D)
sensor.
A
sensor
was
used
collect
3D
point
cloud
data
from
orchard.
Six
samples
trees,
representing
variety
shapes
sizes,
were
selected
collection
validation.
Commercial
software
python
programming
language
utilized
process
collected
data.
processing
steps
involved
conversion,
radius
outlier
removal,
voxel
grid
downsampling,
denoising
through
filtering
erroneous
points,
segmentation
region
interest
(ROI),
clustering
density-based
spatial
(DBSCAN)
algorithm,
transformation,
removal
ground
points.
Accuracy
assessed
by
comparing
estimated
outputs
with
corresponding
measured
values.
sensor-estimated
heights
3.05
±
0.34
m
3.13
0.33
m,
respectively,
mean
absolute
error
(MAE)
0.08
root
squared
(RMSE)
0.09
linear
coefficient
determination
(r2)
0.98,
confidence
interval
(CI)
−0.14
−0.02
high
concordance
correlation
(CCC)
0.96,
indicating
strong
agreement
accuracy.
volumes
13.76
2.46
m3
14.09
2.10
m3,
MAE
0.57
RMSE
0.61
r2
value
0.97,
CI
−0.92
0.26,
demonstrating
precision.
For
distances
3.04
0.17
3.18
0.24
3.35
3.40
0.05
values
0.12
0.92
0.07
0.94
respectively.
−0.18
0.01,
−0.1,
0.002
Although
minor
differences
observed,
estimates
efficient,
though
specific
measurements
require
further
refinement.
results
are
based
on
limited
dataset
six
values,
providing
initial
insights
into
performance.
However,
larger
would
offer
more
reliable
accuracy
assessment.
small
sample
size
(six
trees)
limits
generalizability
findings
necessitates
caution
interpreting
results.
Future
studies
should
incorporate
broader
diverse
validate
refine
characterization,
enhancing
practices
