As
long-distance
oil
pipelines
near
the
end
of
their
operational
tenure,
propensity
for
leakage
due
to
localized
defects
markedly
increases,
necessitating
imperative
systematic
inspection
and
sustained
maintenance
efforts.
Magnetic
Flux
Leakage
(MFL)
inspection,
a
mainstream
non-destructive
testing
methodology,
has
been
extensively
adopted.
In
light
voluminous
nature
monitoring
data,
deep
learning
computer
vision
technologies
play
pivotal
role
in
enhancing
efficiency
accuracy
detection.
This
study
introduces
an
innovative
cascading
detection
technique
that
amalgamates
advanced
visual
recognition
network
YOLOv8
with
novel
multi-input
parallel
convolution
structure.
Through
channel
fusion-based
image
preprocessing
techniques,
it
adeptly
utilizes
tri-axial
MFL
experimental
data
precisely
localize
pipeline
defects,
while
concurrently
predicting
sizes
depths
defects.
research
meticulously
investigates
impact
various
processing
techniques
model
architectures
on
defect
quantifiable
prediction.
Following
stringent
validation,
our
method
demonstrated
superiority
over
conventional
approaches
quantitative
assessment
tasks.
Moreover,
proposed
significantly
outperforms
single-input
prediction
networks
predictive
concerning
highlighting
its
prospective
utility
gas
through
improved
precision,
timeliness,
economic
interventions.
