Urban Climate,
Journal Year:
2023,
Volume and Issue:
49, P. 101470 - 101470
Published: March 17, 2023
Cool
Pavements
(CPs)
can
maintain
a
lower
surface
temperature
than
conventional
pavements
and
mitigate
urban
overheating.
CPs
decrease
their
heat
gains
by
enhancing
pavement’s
radiative
properties,
i.e.
solar
reflectance
thermal
emissivity,
performing
evaporating
cooling,
or
converting
to
other
forms
of
energy.
Several
studies
have
reported
substantial
decreases,
however,
wide
application
is
still
impeded.
Most
the
CP
report
on
in-lab
investigations
numerical
evaluations,
while
only
few
performance
under
real-life
boundary
conditions.
This
review
reports
performed
in
outdoors
with
respect
reflective,
evaporative
energy
storage
techniques.
The
corresponding
protocols
are
analyzed
for
various
scales
evaluation
critically
discussed
limitations,
research
gaps
future
paths.
Also,
monitoring
protocol
proposed
outdoor
CPs.
analysis
showed
that
there
lack
relevant
standards,
whilst
cooling
effects
vary
within
3–20
°C,
8–25
4–14
4–19
°C
permeable,
storage,
large-scale
applications,
respectively.
IEEE Transactions on Geoscience and Remote Sensing,
Journal Year:
2020,
Volume and Issue:
59(9), P. 7844 - 7853
Published: Nov. 6, 2020
Few-shot
scene
classification
aims
to
recognize
unseen
concepts
from
few
labeled
samples.
However,
most
existing
works
are
generally
inclined
learn
metalearners
or
transfer
knowledge
while
ignoring
the
importance
discriminative
representations
and
a
proper
metric
for
remote
sensing
images.
To
address
these
challenges,
in
this
article,
we
propose
an
end-to-end
network
boosting
few-shot
image
classification,
called
learning
of
adaptive
match
(DLA-MatchNet).
Specifically,
first
adopt
attention
technique
delve
into
interchannel
interspatial
relationships
automatically
discover
regions.
Then,
channel
spatial
modules
can
be
incorporated
with
feature
by
using
different
fusion
schemes,
achieving
"discriminative
learning."
Afterward,
considering
issues
large
intraclass
variances
interclass
similarity
images,
instead
simply
computing
distances
between
support
samples
query
samples,
concatenate
features
depth
utilize
matcher
"adaptively"
select
semantically
relevant
sample
pairs
assign
scores.
Our
method
leverages
episode-based
strategy
train
model.
Once
trained,
our
model
predict
category
without
further
fine-tuning.
Experimental
results
on
three
public
data
sets
demonstrate
effectiveness
task.
IEEE Geoscience and Remote Sensing Magazine,
Journal Year:
2021,
Volume and Issue:
9(4), P. 68 - 101
Published: April 5, 2021
Change
detection
is
a
vibrant
area
of
research
in
remote
sensing.
Thanks
to
increases
the
spatial
resolution
sensing
images,
subtle
changes
at
finer
geometrical
scale
can
now
be
effectively
detected.
However,
change
from
very-high-spatial-resolution
(VHR)
(≤5
m)
images
challenging
due
limited
spectral
information,
variability,
geometric
distortion,
and
information
loss.
To
address
these
challenges,
many
algorithms
have
been
developed.
comprehensive
review
VHR
lacking
existing
literature.
This
aims
fill
gap
mainly
includes
three
aspects:
methods,
applications,
future
directions.
