Processes,
Год журнала:
2024,
Номер
12(6), С. 1075 - 1075
Опубликована: Май 24, 2024
Sustainable
and
green
waste
management
has
become
increasingly
crucial
due
to
the
rising
volume
of
driven
by
urbanization
population
growth.
Deep
learning
models
based
on
image
recognition
offer
potential
for
advanced
classification
recycling
methods.
However,
traditional
approaches
usually
rely
single-label
images,
neglecting
complexity
real-world
occurrences.
Moreover,
there
is
a
scarcity
efforts
directed
at
actual
municipal
data,
with
most
studies
confined
laboratory
settings.
Therefore,
we
introduce
an
efficient
Query2Label
(Q2L)
framework,
powered
Vision
Transformer
(ViT-B/16)
as
its
backbone
complemented
innovative
asymmetric
loss
function,
designed
effectively
handle
multi-label
classification.
Our
experiments
newly
developed
dataset
“Garbage
In,
Garbage
Out”,
which
includes
25,000
street-level
each
potentially
containing
up
four
types
waste,
showcase
Q2L
framework’s
exceptional
ability
identify
accuracy
exceeding
92.36%.
Comprehensive
ablation
experiments,
comparing
different
backbones,
functions,
substantiate
efficacy
our
approach.
model
achieves
superior
performance
compared
models,
mean
average
precision
increase
2.39%
when
utilizing
switching
ViT-B/16
improves
4.75%
over
ResNet-101.
Organizacija,
Год журнала:
2025,
Номер
58(1), С. 3 - 19
Опубликована: Фев. 1, 2025
Abstract
Background
and
Purpose
The
purpose
of
this
study
is
to
investigate
how
various
types
innovation
impact
sustainability
measures
within
manufacturing
companies;
these
include
minimizing
raw
material
usage,
reducing
energy
consumption,
optimizing
waste
management.
research
further
evaluates
the
linkage
between
job
creation,
focusing
on
fosters
new
employment
opportunities
enhances
in
sector.
Methodology
methodology
involves
a
hierarchical
regression
analysis
conducted
sample
1,570
companies
Colombia
using
SPSS
software.
This
approach
aims
quantitatively
assess
effectiveness
innovation,
sustainability,
policies
industrial
organizations.
Results
findings
reveal
significant
insights
into
their
management
environmental
sustainability.
These
results
underscore
practical
implications
embracing
for
long-term
benefits,
despite
immediate
costs.
Conclusion
provides
comprehensive
examination
diverse
consequential
impacts
Additionally,
it
suggests
directions
future
that
could
refine
enhance
practices
industry.
FUDMA Journal of Sciences,
Год журнала:
2025,
Номер
9(2), С. 193 - 208
Опубликована: Фев. 28, 2025
Tannery
effluent
poses
significant
risks
to
soil
health,
primarily
through
contamination
with
heavy
metals
like
chromium,
sulphides,
and
persistent
organic
pollutants
(POPs).
These
toxic
substances
inhibit
microbial
activity,
reducing
nutrient
cycling
matter
decomposition
essential
for
fertility.
Beneficial
microorganisms,
including
nitrogen-fixing
bacteria,
are
particularly
affected,
leading
altered
communities
dominated
by
less
advantageous,
metal-tolerant
species.
Accumulation
of
POPs
disrupts
enzymatic
activities,
interferes
plant
root
growth,
complicates
remediation
efforts
due
pollutant
migration
groundwater
potential
entry
into
the
food
chain.
Prolonged
exposure
such
contaminants
diminishes
fertility,
reduces
resilience,
ecosystem
services,
posing
threats
agricultural
productivity
environmental
health.
This
review
was
aimed
outline
what
made
bioremediation
a
superior
treatment
technology
among
other
methods
used
in
remediating
tannery
contaminated
soil.
Efforts
mitigate
impacts
involve
combination
physical,
chemical,
biological
technologies.
Physical
washing,
flushing,
thermal
desorption
focus
on
removing
or
isolating
contaminants,
while
chemical
approaches
as
oxidation,
reduction,
stabilization
transform
harmful
forms
immobilize
them.
Biological
leverages
microorganisms
plants
detoxify
sustainably.
Bioremediation
strategies
aid
bioaugmentation
biostimulation
do
enhance
activity
address
inorganic
effectively
more
than
physical
methods.
Another
excellent
called
phytoremediation
can
also
effectively,
Achieving
better
technique
should
be
coupled
stringent
industrial
regulations,
sustainable
tanning
methods,
stakeholder
awareness
IOP Conference Series Earth and Environmental Science,
Год журнала:
2025,
Номер
1452(1), С. 012034 - 012034
Опубликована: Фев. 1, 2025
Abstract
The
growing
volume
of
waste
presents
substantial
management
issues,
especially
under
the
current
collect-transport-dispose
paradigm,
which
frequently
results
in
overburdened
temporary
disposal
sites
(TDS
or
TPS)
Malang
Regency
since
there
is
no
reduction
prior
to
TDS.
Organic
waste,
as
significant
fraction
was
focus
study.
Mass
Balance
Analysis
used
calculate
Recovery
Factor
(RF)
TDS
for
calculating
potential
four
Pakis
Sub-district,
through
scenario.
Two
scenarios
were
study
focusing
on
increase,
i.e.
maintaining
RF
and
changing
first
second
scenario
respectively.
showed
that
expanding
service
area
still
have
same
capacity
though
increasing
input.
Meanwhile,
scenario,
increases
RF,
can
decrease
36%
transported
into
These
highlight
effectiveness
maggot-based
organic
processing
achieving
targets,
offering
a
scalable
model
sustainable
management.
Processes,
Год журнала:
2024,
Номер
12(6), С. 1075 - 1075
Опубликована: Май 24, 2024
Sustainable
and
green
waste
management
has
become
increasingly
crucial
due
to
the
rising
volume
of
driven
by
urbanization
population
growth.
Deep
learning
models
based
on
image
recognition
offer
potential
for
advanced
classification
recycling
methods.
However,
traditional
approaches
usually
rely
single-label
images,
neglecting
complexity
real-world
occurrences.
Moreover,
there
is
a
scarcity
efforts
directed
at
actual
municipal
data,
with
most
studies
confined
laboratory
settings.
Therefore,
we
introduce
an
efficient
Query2Label
(Q2L)
framework,
powered
Vision
Transformer
(ViT-B/16)
as
its
backbone
complemented
innovative
asymmetric
loss
function,
designed
effectively
handle
multi-label
classification.
Our
experiments
newly
developed
dataset
“Garbage
In,
Garbage
Out”,
which
includes
25,000
street-level
each
potentially
containing
up
four
types
waste,
showcase
Q2L
framework’s
exceptional
ability
identify
accuracy
exceeding
92.36%.
Comprehensive
ablation
experiments,
comparing
different
backbones,
functions,
substantiate
efficacy
our
approach.
model
achieves
superior
performance
compared
models,
mean
average
precision
increase
2.39%
when
utilizing
switching
ViT-B/16
improves
4.75%
over
ResNet-101.