Journal of Water and Health,
Год журнала:
2022,
Номер
20(9), С. 1343 - 1363
Опубликована: Авг. 11, 2022
Accelerated
mining
activities
have
increased
water
contamination
with
potentially
toxic
elements
(PTEs)
and
their
associated
human
health
risk
in
developing
countries.
The
current
study
investigated
the
distribution
of
PTEs,
potential
sources
assessment
both
ground
surface
non-mining
areas
Khyber
Pakhtunkhwa,
Pakistan.
Water
samples
(n
=
150)
were
taken
from
selected
sites
analyzed
for
six
PTEs
(Ni,
Cr,
Zn,
Cu,
Pb
Mn).
Among
Cr
showed
a
high
mean
concentration
(497)
μg
L-1,
followed
by
Zn
(414)
L-1
area,
while
lowest
value
(4.44)
areas.
Elevated
concentrations
Ni,
moderate
level
Mohmand
District
exceeded
permissible
limits
set
WHO.
Multivariate
statistical
analyses
that
pollution
mainly
mafic-ultramafic
rocks,
acid
mine
drainage,
open
dumping
wastes
tailings.
hazard
quotient
(HQ)
was
highest
children
relative
to
adults,
but
not
higher
than
USEPA
limits.
index
(HI)
ingestions
all
lower
threshold
(HIing
<
1),
except
District,
which
HI
>1
through
ingestion.
Moreover,
carcinogenic
(CR)
values
Ni
(1.0E-04-1.0E-06).
In
order
protect
drinking
further
contamination,
management
techniques
policy
operations
need
be
implemented.
Sensors,
Год журнала:
2022,
Номер
22(2), С. 464 - 464
Опубликована: Янв. 8, 2022
In-flight
system
failure
is
one
of
the
major
safety
concerns
in
operation
unmanned
aerial
vehicles
(UAVs)
urban
environments.
To
address
this
concern,
a
framework
consisting
following
three
main
tasks
can
be
utilized:
(1)
Monitoring
health
UAV
and
detecting
failures,
(2)
Finding
potential
safe
landing
spots
case
critical
detected
step
1,
(3)
Steering
to
spot
found
2.
In
paper,
we
specifically
look
at
second
task,
where
investigate
feasibility
utilizing
object
detection
methods
suffers
an
in-flight
failure.
Particularly,
different
versions
YOLO
objection
method
compare
their
performances
for
specific
application
location
that
has
suffered
We
performance
YOLOv3,
YOLOv4,
YOLOv5l
while
training
them
by
large
image
dataset
called
DOTA
Personal
Computer
(PC)
also
Companion
(CC).
plan
use
chosen
algorithm
on
CC
attached
UAV,
PC
used
verify
trends
see
between
algorithms
CC.
confirm
these
effective
emergency
report
accuracy
speed
application.
Our
investigation
shows
outperforms
YOLOv4
YOLOv3
terms
maintaining
slightly
slower
inference
speed.
Engineering Applications of Computational Fluid Mechanics,
Год журнала:
2018,
Номер
12(1), С. 810 - 823
Опубликована: Янв. 1, 2018
This
study
explores
the
river-flow-induced
impacts
on
performance
of
machine
learning
models
applied
for
forecasting
water
quality
parameters
in
coastal
waters
Hilo
Bay,
Pacific
Ocean.
For
this
purpose,
hourly
recorded
salinity,
temperature
and
turbidity
as
well
flow
data
Wailuku
River
were
used.
Several
including
artificial
neural
network,
extreme
support
vector
regression
have
been
employed
to
investigate
impact
from
current
time
up
2
h
ahead.
Following
input
structure
models,
two
separate
based
excluding
river
developed
each
variable
quantify
importance
discharge
accuracy
models.
The
different
was
found
be
close
other
showing
similar
pattern
considering
uncertainty
forecasts.
results
revealed
that
influenced
salinity
bay
which
variables
had
better
compared
with
those
series.
Among
investigated
research,
made
most
least
improvement
efficiency
temperature,
respectively.
Overall,
it
observed
can
properly
forecasted
several
hours
ahead
providing
a
potentially
valuable
tool
environmental
management
monitoring
areas.
