I
A
E
S
I
n
t
e
r
n
at
io
n
al
Jou
r
n
al
of
A
r
t
if
ic
ia
l
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
V
ol
.
14
, N
o.
3
,
J
une
20
25
, pp.
2282
~
2290
I
S
S
N
:
2252
-
8938
,
D
O
I
:
10.11591/
ij
a
i.
v
14
.i
3
.pp
2282
-
2290
2282
Jou
r
n
al
h
om
e
page
:
ht
tp
:
//
ij
ai
.
ia
e
s
c
or
e
.c
om
R
oof
t
op
s d
e
t
e
c
t
i
on
w
i
t
h
Y
O
L
O
v8 f
r
om
ae
r
i
al
i
m
age
r
y an
d
a
b
r
i
e
f
r
e
vi
e
w
on
r
oof
t
op
p
h
ot
ovol
t
ai
c
p
ot
e
n
t
i
al
ass
e
ss
m
e
n
t
M
d
. S
ab
b
ir
A
h
m
e
d
1
, M
d
. S
h
oh
e
l
A
r
m
an
1
, N
u
s
r
at
T
as
n
im
1
,
M
d
H
af
iz
u
l
I
m
r
an
1
, M
u
s
ab
b
ir
H
as
an
S
am
m
ak
1
, T
ou
h
id
B
h
u
iy
an
2
1
D
e
pa
r
t
m
e
nt
of
S
of
t
w
a
r
e
E
ngi
ne
e
r
i
ng, D
I
U
D
a
t
a
S
c
i
e
nc
e
L
a
b, D
a
f
f
odi
l
I
nt
e
r
na
t
i
ona
l
U
ni
ve
r
s
i
t
y
,
D
a
f
f
odi
l
S
m
a
r
t
C
i
t
y
,
B
a
ngl
a
de
s
h
2
S
c
hool
of
I
nf
or
m
a
t
i
on T
e
c
hnol
ogy
,
W
a
s
hi
ngt
on U
ni
ve
r
s
i
t
y of
S
c
i
e
nc
e
a
nd T
e
c
hnol
ogy
,
A
l
e
xa
ndr
i
a
,
U
ni
t
e
d S
t
a
t
e
s
A
r
t
ic
le
I
n
f
o
A
B
S
T
R
A
C
T
A
r
ti
c
le
h
is
to
r
y
:
R
e
c
e
iv
e
d
A
ug
26, 2024
R
e
vi
s
e
d
J
a
n 29, 2025
A
c
c
e
pt
e
d
M
a
r
15, 2025
Recent
years
have
seen
significant
advancements
in
the
switch
from
fossil
fuel
-
based
energy
systems
to
renewable
energy.
Decentralized
solar
photovoltaic
(PV)
is
one
of
the
most
promising
energy
sources
since
there
is
a
lot
of
rooftop
space,
it
is
easy
to
install,
and
the
cost
of
the
PV
pa
nels
is
low.
The
determination
of
rooftop
locations
for
PV
installation
is
cru
cial
for
energy
planning.
With
this
context,
this
study
aimed
to
detect
the
s
uitable
rooftops
of
different
shapes.
The
dataset
of
5,076
building
roofs
used
in
this
study
was
gathered
by
us
utilizing
a
drone.
This
study
identifi
ed
ten
distinct
roof
shapes
accurately,
including
triangle,
square,
penta,
hexa,
hepta
,
octa,
nona,
deca,
gabled
roof,
and
hipped
roof
,
using
the
most
recent
vers
ion
of
you
only
live
once
(
YOLO
)
,
known
as
YOLOv8.
Recent
research
re
vealed,
YOLOv8
is
more
accurate
than
earlier
YOLO
models
which
is
the
reason
of
utilizing
YOLOv8.
Accuracy
of
this
work
of
rooftops
detection
is
93.6%.
Also,
the
precision,
recall
,
and
F1
-
score
confidence
curve
showed
good
performances
too.
Finally,
a
brief
review
of
the
most
recent
studies
on
the
evaluatio
n
of
rooftop
PV
potential
was
conducted
to
provide
insight
i
nto
the
use of solar energy.
K
e
y
w
o
r
d
s
:
A
e
r
ia
l
i
m
a
ge
s
D
e
e
p
l
e
a
r
ni
ng
P
V
m
odul
e
s
R
oof
to
ps
Y
O
L
O
v8
This is an
open
acce
ss artic
le unde
r the
CC BY
-
SA
license.
C
or
r
e
s
pon
di
n
g A
u
th
or
:
N
us
r
a
t
T
a
s
ni
m
D
e
pa
r
tm
e
nt
of
S
of
twa
r
e
E
ngi
ne
e
r
in
g, D
I
U
D
a
ta
S
c
ie
nc
e
L
a
b, D
a
f
f
odi
l
I
nt
e
r
na
ti
ona
l
U
ni
ve
r
s
it
y
D
a
f
f
odi
l
S
m
a
r
t
C
it
y, B
ir
ul
ia
-
1216, Dha
ka
, B
a
ngl
a
de
s
h
E
m
a
il
:
nus
r
a
tt
a
s
ni
m
17@
gm
a
il
.c
om
1.
I
N
T
R
O
D
U
C
T
I
O
N
A
s
th
e
de
m
a
nd
f
or
r
e
ne
w
a
bl
e
e
ne
r
gy
r
is
e
s
gl
oba
ll
y,
s
ol
a
r
pow
e
r
ha
s
e
m
e
r
ge
d
a
s
a
pr
a
c
ti
c
a
l
a
nd
s
us
ta
in
a
bl
e
c
hoi
c
e
.
P
hot
ovol
ta
ic
(
P
V
)
m
odul
e
s
,
w
hi
c
h
c
onve
r
t
s
unl
ig
ht
in
to
e
le
c
tr
ic
it
y.
H
ow
e
ve
r
,
lo
c
a
ti
ng
s
ui
ta
bl
e
r
oof
to
p
s
ur
f
a
c
e
s
is
not
a
n
e
a
s
y
ta
s
k.
I
n
th
is
s
tu
dy,
a
u
to
m
a
te
d
de
e
p
le
a
r
ni
ng
a
ppr
oa
c
he
s
ha
ve
be
e
n
in
ve
s
ti
ga
te
d
to
id
e
nt
if
y
a
nd
a
s
s
e
s
s
r
oof
to
p
z
one
s
th
a
t
a
r
e
s
ui
ta
bl
e
f
or
P
V
m
odul
e
s
f
r
o
m
a
e
r
ia
l
im
a
ge
s
.
N
ow
m
or
e
t
ha
n e
ve
r
, pe
opl
e
a
r
e
c
ons
id
e
r
in
g l
ow
-
c
a
r
bon de
ve
lo
pm
e
n
t
a
nd t
he
us
e
of
r
e
ne
w
a
bl
e
e
n
e
r
gy. I
ns
ta
ll
a
ti
on
of
P
V
pa
ne
ls
c
a
n
s
ur
e
ly
r
e
duc
e
a
m
ount
of
gr
e
e
nhous
e
ga
s
e
m
i
s
s
io
ns
.
T
h
e
m
a
jo
r
it
y
of
c
ount
r
ie
s
r
e
ly
on
f
os
s
il
f
ue
l
-
ba
s
e
d
e
ne
r
gy
pr
oduc
ti
on;
how
e
ve
r
,
th
is
m
e
th
od
is
e
xpe
ns
iv
e
.
T
he
hi
gh
c
os
t
of
th
is
ki
nd
of
e
ne
r
gy
pr
oduc
ti
on
is
a
m
a
jo
r
c
onc
e
r
n
f
or
de
ve
lo
pi
ng
c
ount
r
ie
s
.
I
n
c
ount
r
ie
s
w
it
h
s
c
a
r
c
e
na
tu
r
a
l
r
e
s
our
c
e
s
a
nd
in
a
de
qua
te
e
n
e
r
gy
m
a
na
ge
m
e
nt
c
a
p
a
c
it
ie
s
,
pa
r
ti
c
ul
a
r
ly
in
B
a
ngl
a
de
s
h,
th
e
r
e
is
a
s
e
ve
r
e
e
ne
r
gy
s
hor
ta
ge
.
F
r
om
t
hi
s
s
it
ua
ti
on, i
t
is
c
le
a
r
t
ha
t
e
m
pl
oyi
ng s
ol
a
r
pa
ne
ls
i
s
no
w
ve
r
y ne
c
e
s
s
a
r
y.
B
ut
th
e
f
ir
s
t
s
te
p
to
in
s
ta
ll
P
V
m
odul
e
s
on
r
oof
to
ps
is
to
f
in
d
out
th
e
s
ui
ta
bl
e
r
oof
to
ps
.
T
he
s
e
a
r
c
h
f
or
a
c
c
e
pt
a
bl
e
r
oof
to
p
s
ur
f
a
c
e
s
f
or
th
e
in
s
ta
ll
a
ti
on
o
f
P
V
m
od
ul
e
s
,
how
e
ve
r
,
is
ty
pi
c
a
ll
y
a
ti
m
e
-
c
ons
um
in
g,
e
xpe
ns
iv
e
,
a
nd
la
bor
-
in
te
ns
iv
e
pr
oc
e
s
s
th
a
t
r
e
qui
r
e
s
phys
ic
a
l
e
xa
m
in
a
ti
on.
I
n
o
r
de
r
to
m
a
ke
it
s
im
pl
e
to
in
s
ta
ll
P
V
m
odul
e
s
on
th
e
a
ppr
opr
ia
te
r
oof
to
ps
qui
c
kl
y
a
nd
w
it
h
m
in
im
a
l
la
bor
,
th
is
s
tu
dy
a
im
s
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
R
oof
to
ps
de
te
c
ti
on w
it
h Y
O
L
O
v
8 f
r
om
ae
r
ia
l
image
r
y
and a br
i
e
f
r
e
v
ie
w
on r
oof
to
p …
(
M
d. Sabbir
A
hm
e
d)
2283
a
ut
om
a
ti
c
a
ll
y
de
te
c
t
r
oof
to
ps
a
nd
c
la
s
s
if
y
th
e
m
in
to
va
r
io
us
c
a
te
gor
ie
s
.
T
hi
s
s
tu
dy
u
s
e
s
a
da
ta
s
e
t
th
a
t
w
e
ha
d
c
ol
le
c
te
d
on
our
ow
n.
W
e
c
a
pt
ur
e
d
th
e
a
e
r
ia
l
f
oot
a
ge
by
us
in
g
a
dr
one
to
c
a
pt
ur
e
a
e
r
ia
l
im
a
ge
s
.
A
e
r
ia
l
phot
ogr
a
phs
s
hot
by
a
dr
one
pr
ovi
de
a
bi
r
d'
s
-
e
ye
pe
r
s
pe
c
ti
ve
of
bui
ld
in
gs
a
nd
la
nds
c
a
pe
s
.
A
e
r
ia
l
im
a
gi
ng
gi
ve
s
s
p
a
ti
a
l
da
ta
th
a
t
c
a
n
be
ut
il
iz
e
d
to
qui
c
kl
y
a
nd
a
c
c
ur
a
t
e
ly
a
na
ly
z
e
la
r
ge
a
r
e
a
s
,
doi
ng
a
w
a
y
w
it
h
th
e
ne
c
e
s
s
it
y
f
or
phys
ic
a
l
in
s
pe
c
ti
on
s
.
W
it
h
a
r
unt
im
e
of
f
iv
e
to
s
e
ve
n
m
in
ut
e
s
e
a
c
h,
w
e
r
e
c
or
de
d
a
to
ta
l
of
30 f
il
m
s
. T
he
vi
de
os
w
e
r
e
t
he
n c
onve
r
te
d i
nt
o i
m
a
ge
s
.
A
f
te
r
r
e
vi
e
w
in
g
e
a
r
li
e
r
s
tu
di
e
s
[
1]
–
[
4]
th
a
t
in
di
c
a
te
d
a
gr
e
a
te
r
a
c
c
ur
a
c
y
of
de
t
e
c
ti
on
w
he
n
ut
il
iz
in
g
th
e
Y
O
L
O
v8
m
ode
l,
w
e
m
a
de
th
e
de
c
i
s
io
n
to
e
m
pl
oy
it
.
Y
ou
onl
y
lo
ok
onc
e
(
Y
O
L
O
)
is
a
n
obj
e
c
t
de
te
c
ti
on
a
lg
or
it
hm
.
I
t
is
a
popula
r
r
e
a
l
-
ti
m
e
obj
e
c
t
de
t
e
c
ti
on
s
y
s
te
m
th
a
t
c
a
n
id
e
nt
if
y
a
nd
lo
c
a
li
z
e
m
ul
ti
pl
e
obj
e
c
t
s
in
a
n
im
a
ge
or
vi
de
o
s
tr
e
a
m
.
Y
O
L
O
v8
is
th
e
m
os
t
upda
te
d
ve
r
s
io
n
of
Y
O
L
O
r
ig
ht
now
.
R
e
c
e
nt
r
e
s
e
a
r
c
h
r
e
ve
a
le
d
th
a
t
Y
O
L
O
v8
is
m
or
e
a
c
c
ur
a
te
th
a
n
e
a
r
li
e
r
Y
O
L
O
m
ode
ls
.
A
ddi
ti
ona
ll
y,
Y
O
L
O
v8'
s
tr
a
in
in
g
ti
m
e
w
il
l
li
ke
ly
be
s
hor
te
r
t
ha
n t
ha
t
of
pr
e
vi
ous
t
w
o
-
s
ta
ge
obj
e
c
t
de
te
c
ti
on mode
ls
[
5]
.
F
ol
lo
w
in
g t
he
r
oof
to
p
de
te
c
ti
on pha
s
e
, w
e
c
onduc
te
d a
br
ie
f
r
e
vi
e
w
on t
he
ot
he
r
pot
e
nt
ia
li
ti
e
s
of
t
hi
s
r
oof
to
p
de
te
c
ti
on
r
e
s
e
a
r
c
h,
in
c
lu
di
ng
e
s
ti
m
a
ti
ng
m
a
th
e
m
a
ti
c
a
ll
y
th
e
num
be
r
of
P
V
m
odul
e
s
th
a
t
c
a
n
be
in
s
ta
ll
e
d
on
a
pa
r
ti
c
ul
a
r
r
oof
to
p,
de
te
r
m
in
in
g
th
e
s
h
a
pe
a
nd
a
ngl
e
of
th
e
r
oof
to
p
th
a
t
is
be
s
t
s
ui
te
d
f
or
th
e
hi
ghe
s
t
s
ol
a
r
pow
e
r
ge
ne
r
a
ti
on,
a
nd e
s
ti
m
a
ti
on
of
s
ol
a
r
ir
r
a
di
a
n
c
e
.
S
o,
th
e
c
ont
r
ib
ut
io
ns
of
th
is
r
e
s
e
a
r
c
h
w
or
k
c
a
n be
s
um
m
a
r
iz
e
d a
s
f
ol
lo
w
s
:
−
C
r
e
a
ti
ng a
da
ta
s
e
t
of
dr
one
c
a
pt
ur
e
d a
r
ia
l
im
a
ge
s
w
it
h hi
gh r
e
s
o
lu
ti
on t
o ge
t
ve
r
y a
c
c
ur
a
te
de
te
c
ti
on.
−
U
ti
li
z
in
g
th
e
m
os
t
r
e
c
e
nt
Y
O
L
O
v8
m
ode
l
f
or
de
te
c
ti
on
a
n
d,
f
ur
th
e
r
pot
e
nt
ia
li
ty
a
na
ly
s
is
of
r
oof
to
p
de
te
c
ti
on f
or
P
V
i
ns
ta
ll
a
ti
on.
2.
L
I
T
E
R
A
T
U
R
E
R
E
V
I
E
W
Z
hong
e
t
al
.
[
6]
pr
opos
e
d
a
f
r
a
m
e
w
or
k
us
in
g
hi
gh
-
r
e
s
ol
ut
io
n
s
a
te
ll
it
e
im
a
ge
s
a
va
il
a
bl
e
e
m
pl
oyi
ng
a
de
e
p
le
a
r
ni
ng
-
ba
s
e
d
te
c
hni
que
f
or
a
ut
om
a
ti
c
a
ll
y
e
xt
r
a
c
ti
ng
r
oo
f
to
p
a
r
e
a
s
.
T
he
e
s
ti
m
a
te
d
r
oof
to
p
a
r
e
a
s
ui
ta
bl
e
f
or
P
V
in
s
ta
ll
a
ti
ons
in
N
a
nj
in
g
w
a
s
f
ound
to
be
330.36
km
2
,
w
it
h
a
n
im
pr
e
s
s
iv
e
ove
r
a
ll
a
c
c
ur
a
c
y
of
0.92
.
M
a
o
e
t
al
.
[
7]
r
e
vi
e
w
s
va
r
io
us
m
e
th
ods
f
or
id
e
nt
if
yi
ng
P
V
in
s
ta
ll
a
ti
ons
,
a
nd
pr
opos
e
s
opt
im
iz
a
ti
ons
to
e
nha
nc
e
th
e
id
e
nt
if
ic
a
ti
on
pr
oc
e
s
s
a
nd
f
or
e
c
a
s
t
r
oof
to
p
P
V
pot
e
nt
ia
l.
D
e
e
p
le
a
r
ni
ng,
e
xhi
bi
ts
s
upe
r
io
r
a
c
c
ur
a
c
y
in
s
e
gm
e
nt
in
g
P
V
s
y
s
te
m
s
of
a
ll
s
i
z
e
s
,
w
it
h
r
oof
to
p
P
V
s
e
gm
e
nt
a
ti
on
a
c
hi
e
vi
ng
pr
e
c
is
io
n a
nd
r
e
c
a
ll
r
a
te
s
r
a
ngi
ng
f
r
om
41
to
98.9%
a
nd
54.5
to
95.8%
,
r
e
s
pe
c
ti
ve
ly
.
A
s
la
ni
a
nd
S
e
ip
e
l
[
8]
a
im
s
to
pr
e
s
e
nt
a
c
om
pr
e
he
ns
iv
e
m
e
th
odol
ogy
th
a
t
in
c
lu
de
s
a
ut
om
a
ti
c
e
xt
r
a
c
ti
on
of
bui
ld
in
g
f
oot
p
r
in
ts
,
s
e
gm
e
nt
a
ti
on
of
r
oo
f
f
a
c
e
s
,
a
nd
id
e
nt
if
ic
a
ti
on
of
s
ui
ta
bl
e
r
oof
to
p
a
r
e
a
s
f
or
s
ol
a
r
in
f
r
a
s
tr
uc
tu
r
e
.
T
he
e
xpe
r
im
e
nt
a
l
r
e
s
ul
ts
de
m
ons
tr
a
te
d
im
pr
e
s
s
iv
e
a
c
c
ur
a
c
y
of
95%
in
bui
ld
in
g
e
xt
r
a
c
ti
on.
M
oha
je
r
i
e
t
al
.
[
9]
a
im
s
to
e
nha
nc
e
s
ol
a
r
e
ne
r
gy
de
pl
oym
e
nt
in
ur
ba
n
a
r
e
a
s
us
in
g
P
V
in
s
ta
ll
a
ti
ons
.
T
he
r
e
s
e
a
r
c
he
r
s
ut
il
iz
e
d
a
m
a
c
hi
ne
le
a
r
ni
ng
te
c
hni
que
c
a
ll
e
d
s
uppor
t
ve
c
to
r
m
a
c
hi
ne
(
S
V
M
)
c
la
s
s
if
ic
a
ti
on
to
c
la
s
s
if
y
10,085
bui
ld
in
g
r
oof
s
in
G
e
ne
va
ba
s
e
d
on
th
e
ir
s
ol
a
r
e
ne
r
gy
pot
e
nt
ia
l.
T
he
S
V
M
a
c
hi
e
ve
d
a
66
%
a
c
c
ur
a
c
y
in
id
e
nt
if
yi
ng
s
ix
r
oof
s
ha
pe
ty
pe
s
.
S
ong
e
t
al
.
[
10]
f
oc
us
e
s
on
ut
il
iz
in
g
s
ol
a
r
e
ne
r
gy
th
r
ough
th
e
in
s
ta
ll
a
ti
on
of
s
ol
a
r
P
V
s
ys
te
m
s
on
bui
ld
in
g
r
oof
to
ps
.
R
e
s
e
a
r
c
h
by
L
e
e
e
t
al
.
[
11]
,
a
nove
l
da
ta
-
dr
iv
e
n
a
pp
r
oa
c
h
to
a
s
s
e
s
s
th
e
s
ol
a
r
pot
e
nt
ia
l
of
r
oof
to
ps
us
in
g
w
id
e
ly
a
v
a
il
a
bl
e
s
a
te
ll
it
e
im
a
ge
s
. T
he
a
ppr
oa
c
h
w
a
s
th
or
oughly
e
va
lu
a
te
d
on
a
n
a
nnot
a
te
d
r
oof
da
ta
s
e
t,
va
li
da
te
d
by
s
ol
a
r
e
xpe
r
ts
,
a
nd
c
om
pa
r
e
d
to
a
li
ght
de
te
c
ti
on
a
n
d
r
a
ngi
ng
(
L
I
D
A
R
)
-
ba
s
e
d
m
e
th
od.
D
e
e
pR
oof
de
m
ons
tr
a
te
d hi
gh a
c
c
ur
a
c
y i
n e
xt
r
a
c
ti
ng r
oof
ge
om
e
tr
y, a
c
hi
e
v
in
g a
t
r
ue
pos
it
iv
e
r
a
te
of
91.1%
.
R
e
s
e
a
r
c
h
by
Z
h
ong
e
t
al
.
[
12]
,
a
c
om
p
ut
a
ti
ona
l
s
ys
t
e
m
t
ha
t
u
s
e
s
d
e
e
p
l
e
a
r
ni
n
g
to
id
e
nt
if
y
pl
a
nne
d
noi
s
e
ba
r
r
i
e
r
s
it
e
s
b
a
s
e
d
o
n
lo
c
a
l
pol
ic
i
e
s
a
n
d
r
e
c
ogni
z
e
c
ur
r
e
n
t
noi
s
e
b
a
r
r
ie
r
s
it
e
s
f
r
om
s
tr
e
e
t
-
vi
e
w
phot
o
s
i
s
pr
opos
e
d
i
n
or
d
e
r
to
e
s
ti
m
a
t
e
th
e
s
ol
a
r
P
V
p
ot
e
nt
ia
l
in
c
it
ie
s
. T
h
e
s
e
r
e
s
ul
t
s
d
e
m
on
s
tr
a
t
e
th
e
s
y
s
te
m
s
'
e
nor
m
ou
s
pot
e
nt
i
a
l
to
s
uppor
t
ur
b
a
n
r
e
n
e
w
a
bl
e
e
ne
r
g
y
s
our
c
e
s
.
A
r
e
a
l
-
ti
m
e
m
ul
ti
va
r
i
a
nt
d
e
e
p
l
e
a
r
ni
n
g
m
ode
l
(
R
M
V
D
M
)
is
s
ugg
e
s
t
e
d
in
[
13]
a
s
a
m
e
t
hod
f
or
id
e
nt
if
yi
ng
a
nd
c
a
te
gor
iz
in
g
P
V
pr
o
bl
e
m
s
.
T
he
s
u
gge
s
te
d
R
M
V
D
M
pe
r
f
or
m
s
w
e
ll
, r
e
a
c
hi
ng
a
n a
c
c
ur
a
c
y
of
a
b
out
97%
.
T
hr
ough id
e
nt
if
yi
n
g t
h
e
t
il
t
a
ngl
e
a
nd
put
ti
n
g
PV
pa
ne
l
s
i
n
th
e
pr
op
e
r
or
ie
nt
a
ti
on,
M
e
m
a
r
i
e
t
al
.
[
14]
in
t
e
nd
s
to
i
m
pr
ove
t
he
a
c
c
ur
a
c
y
of
r
e
a
l
-
ti
m
e
s
ol
a
r
pow
e
r
ge
ne
r
a
ti
on
e
s
ti
m
a
ti
on
in
va
r
io
us
gl
ob
a
l
r
e
gi
o
ns
.
C
he
n
e
t
al
.
[
1
5]
in
tr
od
uc
e
s
a
nov
e
l
a
ppr
o
a
c
h
f
or
id
e
n
ti
f
yi
ng
th
e
s
p
a
ti
a
l
di
s
tr
ib
ut
i
on of
s
ol
a
r
po
w
e
r
pl
a
nt
s
in
l
a
r
g
e
-
s
c
a
le
a
r
e
a
s
.
I
t
d
e
te
c
te
d
52
s
ol
a
r
po
w
e
r
pl
a
nt
s
w
it
h
a
r
e
c
a
ll
r
a
te
of
96.3
0%
.
A
c
c
or
di
ng t
o
T
e
ll
a
e
t
al
.
[
16]
, va
r
io
us
d
e
e
p l
e
a
r
ni
ng ne
t
w
or
ks
w
e
r
e
u
s
e
d t
o c
a
t
e
gor
i
z
e
f
a
u
lt
s
i
n
s
ol
a
r
P
V
c
e
ll
s
. F
r
om
56.29
6%
on t
he
e
lp
v
b
e
n
c
hm
a
r
k
t
o 91
.399
%
on t
he
e
xt
r
a
c
t
e
d
e
lp
v
d
a
ta
s
e
ts
.
T
he
pur
pos
e
of
F
a
khr
a
ia
n
e
t
al
.
[
17]
is
to
c
onduc
t
a
th
o
r
ou
gh,
s
ys
te
m
a
ti
c
r
e
vi
e
w
of
th
e
va
r
io
us
de
ve
lo
pe
d
m
e
th
odol
ogi
e
s
,
id
e
nt
if
y
ke
y
e
le
m
e
nt
s
f
or
e
va
lu
a
ti
ng
ur
ba
n
r
oo
f
to
p
s
ol
a
r
PV
pot
e
nt
ia
l.
I
n
or
de
r
to
id
e
nt
if
y
s
ui
ta
bl
e
r
oof
to
p
s
pa
c
e
s
,
C
a
s
te
ll
o
e
t
al
.
[
18]
us
e
s
a
s
ta
nda
lo
ne
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
k
(
C
N
N
)
,
w
hi
c
h
a
c
hi
e
ve
s
a
n
in
te
r
s
e
c
ti
on
ove
r
uni
on
of
64
%
a
nd
a
n
a
c
c
u
r
a
c
y
of
93%
.
T
he
P
V
s
ys
te
m
de
s
ig
n
th
a
t
w
a
s
pl
a
c
e
d
on
a
n
I
ndi
a
n
r
oof
to
p
is
th
e
s
ubj
e
c
t
of
th
is
r
e
s
e
a
r
c
h
[
19]
.
C
a
de
i
e
t
al
.
[
20]
e
m
pl
oys
s
a
te
ll
it
e
phot
ogr
a
phy
a
nd
m
a
c
hi
ne
le
a
r
ni
ng
te
c
hni
que
s
to
m
a
p
pr
os
pe
c
t
iv
e
r
oof
to
p
s
ur
f
a
c
e
s
f
or
th
e
in
s
ta
ll
a
ti
on
of
P
V
pa
ne
ls
.
I
n
or
de
r
to
e
va
lu
a
te
th
e
r
oof
to
p
a
r
e
a
,
W
ig
in
to
n
e
t
al
.
[
21]
pr
ovi
de
s
a
ppr
oa
c
he
s
th
a
t
in
te
gr
a
te
ge
ogr
a
phi
c
in
f
or
m
a
ti
on
s
ys
te
m
s
a
nd
obj
e
c
t
-
s
pe
c
if
ic
im
a
ge
r
e
c
ogni
ti
on.
T
he
goa
l
of
th
e
s
tu
dy
[
22
]
is
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
3
,
J
une
20
25
:
2282
-
2290
2284
pr
e
s
e
nt
a
th
or
ough
a
na
ly
s
is
of
ge
ogr
a
phi
c
in
f
or
m
a
ti
on
s
ys
te
m
(
G
I
S
)
-
ba
s
e
d
r
oof
to
p
s
ol
a
r
PV
po
te
nt
ia
l
c
a
lc
ul
a
ti
on
te
c
hni
que
s
u
s
e
d
a
t
va
r
io
us
s
iz
e
s
,
in
c
lu
di
ng
na
ti
o
na
l
le
ve
ls
.
T
he
r
e
s
ul
ts
s
how
th
a
t
la
r
ge
-
s
c
a
le
s
pa
ti
a
l
-
te
m
por
a
l
a
s
s
e
s
s
m
e
nt
s
of
f
ut
ur
e
e
ne
r
gy
s
ys
te
m
s
w
it
h
d
e
c
e
nt
r
a
li
z
e
d
e
le
c
tr
ic
a
l
gr
id
s
c
a
n
b
e
pe
r
f
or
m
e
d
us
in
g
e
s
ti
m
a
ti
ng
m
e
th
ods
ba
s
e
d
on
G
I
S
s
.
N
one
of
th
e
r
e
vi
e
w
e
d
s
tu
di
e
s
us
e
d
Y
O
L
O
v8
m
ode
l
ye
t
in
th
is
s
e
c
to
r
. S
o, ut
il
iz
in
g t
hi
s
m
ode
l
a
nd a
na
ly
z
in
g i
ts
r
e
s
ul
t
c
a
n be
a
gr
e
a
t
e
xpe
r
im
e
nt
i
n t
hi
s
f
ie
ld
.
3.
M
E
T
H
O
D
D
a
ta
c
o
ll
e
c
ti
on
w
a
s
done
in
it
ia
ll
y
in
or
de
r
to
lo
c
a
te
s
pe
c
if
ic
bui
l
di
ng
r
oof
t
ops
. T
h
e
d
a
ta
s
e
t
c
ons
i
s
t
s
of
a
e
r
ia
l
phot
ogr
a
p
hs
f
r
om
dr
o
ne
s
th
a
t
w
e
ha
v
e
p
e
r
s
ona
ll
y
a
c
qui
r
e
d
in
th
e
f
or
m
of
vi
d
e
o.
T
he
c
ol
l
e
c
t
e
d
vi
de
o
s
w
e
r
e
t
he
n
t
ur
ne
d
in
to
i
nt
r
o
pi
c
tu
r
e
s
.
T
hr
e
e
pr
e
-
pr
oc
e
s
s
in
g
s
te
ps
-
a
n
not
a
ti
on,
s
c
a
li
ng
,
a
n
d
a
u
gm
e
nt
a
ti
on
-
w
e
r
e
us
e
d
to
pr
e
p
a
r
e
th
e
p
hot
o
s
.
T
h
e
Y
O
L
O
v8
m
ode
l
w
a
s
t
he
n
a
ppl
ie
d
to
th
e
phot
os
.
W
e
u
s
e
Y
O
L
O
v8
,
th
e
m
o
s
t
r
e
c
e
n
t
it
e
r
a
ti
on
of
th
e
Y
O
L
O
m
o
de
l,
w
hi
c
h
c
a
n
be
us
e
d
f
or
ta
s
ks
in
c
lu
di
ng
obj
e
c
t
r
e
c
o
gni
ti
on
,
im
a
ge
c
la
s
s
if
i
c
a
ti
on,
a
nd
in
s
t
a
n
c
e
s
e
gm
e
nt
a
ti
on
.
T
o
s
or
t
th
e
ty
pe
s
of
b
ui
ld
in
g
r
o
of
s
in
th
e
B
a
n
gl
a
d
e
s
hi
m
e
tr
opol
i
s
of
D
ha
k
a
in
to
gr
oup
s
b
a
s
e
d
on
th
e
ir
s
ui
ta
bi
li
ty
f
or
P
V
m
o
dul
e
s
.
R
oof
to
p
s
th
a
t
w
e
r
e
r
e
c
ogni
z
e
d
a
nd
di
s
pl
a
y
e
d
in
s
id
e
boundi
ng
box
e
s
in
th
e
out
put
p
hot
o
s
w
e
r
e
di
vi
de
d
in
t
o
10
di
f
f
e
r
e
nt
ty
p
e
s
.
L
a
s
t
but
n
ot
le
a
s
t,
th
e
li
ke
li
ho
od of
t
hi
s
r
oof
to
p
di
s
c
ov
e
r
y w
a
s
th
e
n
e
x
a
m
in
e
d. F
ig
ur
e
1 s
ho
w
s
t
he
w
or
kf
lo
w
di
a
gr
a
m
.
F
ig
ur
e
1. W
or
kf
lo
w
di
a
gr
a
m
of
r
oof
to
p de
te
c
ti
on
f
or
P
V
i
ns
ta
ll
a
ti
on
3.1
.
D
at
a
c
ol
le
c
t
io
n
T
he
da
ta
s
e
t
ut
il
iz
e
d
in
th
is
r
e
s
e
a
r
c
h
w
or
k
is
our
ow
n
c
ol
le
c
te
d
da
ta
s
e
t.
W
e
c
a
pt
ur
e
d
a
f
e
w
vi
de
o
s
ni
ppe
ts
of
th
e
bui
ld
in
g'
s
r
oof
to
p
us
in
g
dr
one
f
oot
a
ge
.
T
he
s
e
f
il
m
s
w
e
r
e
a
c
qui
r
e
d
by
tr
a
ve
li
ng
to
s
e
ve
r
a
l
c
it
ie
s
’
lo
c
a
le
s
.
W
e
r
e
c
or
de
d
a
to
ta
l
of
30
f
il
m
s
,
w
it
h
a
r
unt
i
m
e
of
be
twe
e
n
f
iv
e
a
nd
s
e
ve
n
m
in
ut
e
s
a
pi
e
c
e
.
T
o obta
in
da
ta
of
t
he
hi
ghe
s
t
qua
li
ty
, c
e
r
ta
in
i
s
s
ue
s
w
e
r
e
pr
e
s
e
r
ve
d dur
in
g da
ta
c
ol
le
c
ti
ng
.
‒
W
e
m
a
de
s
ur
e
a
e
r
ia
l
phot
os
a
r
e
of
a
hi
gh
e
nough
s
ta
nda
r
d
to
pe
r
f
or
m
th
e
de
te
c
ti
ng
w
o
r
k
be
c
a
us
e
im
a
ge
s
w
it
h be
tt
e
r
r
e
s
ol
ut
io
n of
te
n pr
ovi
de
m
or
e
i
nf
or
m
a
ti
on f
or
r
oof
to
p s
tu
dy.
‒
S
in
c
e
not
a
ll
a
r
e
a
s
a
r
e
id
e
a
l
f
or
s
ol
a
r
e
ne
r
gy
in
s
ta
ll
a
ti
on,
th
e
z
one
of
in
te
r
e
s
t
f
or
r
oof
to
p
s
ol
a
r
e
ne
r
gy
in
s
ta
ll
a
ti
ons
ha
s
b
e
e
n c
a
r
e
f
ul
ly
c
hos
e
n f
or
a
e
r
ia
l
im
a
gi
ng.
‒
T
o
e
ns
ur
e
th
a
t
th
e
a
nnot
a
ti
ons
a
r
e
r
e
li
a
bl
e
a
nd
c
or
r
e
c
t,
th
e
da
t
a
s
e
t
w
a
s
c
h
e
c
ke
d
a
f
te
r
it
w
a
s
a
s
s
e
m
bl
e
d.
W
e
pe
r
f
or
m
e
d qua
li
ty
c
he
c
ks
a
nd c
or
r
e
c
ti
on
s
a
s
ne
c
e
s
s
a
r
y.
3.2
.
D
at
as
e
t
d
e
s
c
r
ip
t
io
n
A
f
te
r
c
ol
le
c
ti
ng
vi
de
o
s
hot
in
th
e
a
c
tu
a
l
bui
ld
in
g,
th
e
f
il
m
s
w
e
r
e
tr
a
ns
f
or
m
e
d
in
to
f
r
a
m
e
s
or
phot
ogr
a
phs
. A
f
r
a
m
e
r
a
te
of
one
w
a
s
us
e
d t
o s
a
m
pl
e
t
h
e
vi
de
o
s
. A
bout
500 pic
tu
r
e
s
or
f
r
a
m
e
s
c
oul
d be
m
a
de
by our
t
e
a
m
. W
e
de
le
te
d s
om
e
dupli
c
a
te
photos
a
nd h
a
d 350 f
r
a
m
e
s
l
e
f
t
be
hi
nd.
3.3
.
D
at
a
p
r
e
p
r
oc
e
s
s
in
g
P
r
e
pr
oc
e
s
s
in
g
i
m
a
ge
s
i
s
a
gr
e
a
t
w
a
y t
o e
n
ha
nc
e
t
h
e
ir
qu
a
li
t
y a
nd
pr
e
pa
r
e
t
h
e
m
f
or
a
n
a
ly
s
i
s
a
n
d
f
u
r
th
e
r
pr
o
c
e
s
s
i
ng.
T
hr
oug
h
p
r
e
pr
o
c
e
s
s
i
ng,
w
e
c
a
n
ge
t
r
i
d
o
f
u
nd
e
s
ir
e
d
di
s
t
or
ti
on
s
a
n
d
e
nh
a
n
c
e
c
e
r
ta
in
tr
a
it
s
t
ha
t
a
r
e
c
r
u
c
i
a
l
f
or
r
e
s
e
a
r
c
h
w
or
k
s
.
S
e
v
e
r
a
l
p
r
e
pr
o
c
e
s
s
i
ng
s
te
ps
w
e
r
e
e
x
e
c
ut
e
d f
or
t
he
i
m
a
g
e
s
ut
i
li
z
e
d
i
n t
hi
s
s
t
ud
y.
‒
D
a
ta
a
nnot
a
ti
on:
w
e
id
e
nt
if
ie
d
th
e
r
oof
to
p
r
e
gi
ons
in
th
e
da
t
a
s
e
t
th
a
t
a
r
e
a
ppr
opr
ia
te
f
or
P
V
m
odul
e
in
s
ta
ll
a
ti
on.
T
o
f
ur
th
e
r
le
ve
r
a
ge
th
is
f
o
r
us
a
ge
w
it
h
de
e
p
le
a
r
ni
ng
m
ode
ls
,
th
e
r
oof
to
p
z
one
s
th
a
t
a
r
e
id
e
a
l
f
or
t
he
i
ns
ta
ll
a
ti
on of
s
ol
a
r
e
ne
r
gy s
ys
te
m
s
h
a
s
be
e
n
m
a
nu
a
ll
y r
e
c
ogni
z
e
d.
‒
I
m
a
ge
r
e
-
s
c
a
li
ng:
r
e
s
c
a
li
ng
w
a
s
done
on
th
e
da
t
a
s
e
t
to
pr
ovi
de
c
ons
is
t
e
nc
y
dur
in
g
tr
a
in
in
g
a
nd
to
s
ta
nda
r
di
z
e
t
he
s
i
z
e
of
t
he
a
e
r
ia
l
phot
os
.
‒
D
a
ta
a
ugm
e
nt
a
ti
on:
th
e
da
ta
a
ugm
e
nt
a
ti
on
te
c
hni
qu
e
s
us
e
d
in
th
is
s
tu
dy
w
or
k
in
c
lu
de
r
ot
a
ti
on,
f
li
ppi
ng,
z
oom
in
g,
a
nd
a
dj
us
ti
ng
br
ig
ht
ne
s
s
a
nd
c
ont
r
a
s
t.
T
he
350
f
r
a
m
e
s
w
e
r
e
s
ubs
e
que
nt
ly
in
c
r
e
a
s
e
d
by
560
to
yi
e
ld
t
he
t
r
a
in
, t
e
s
t,
a
nd va
li
da
ti
on da
ta
s
e
ts
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
R
oof
to
ps
de
te
c
ti
on w
it
h Y
O
L
O
v
8 f
r
om
ae
r
ia
l
image
r
y
and a br
i
e
f
r
e
v
ie
w
on r
oof
to
p …
(
M
d. Sabbir
A
hm
e
d)
2285
3.4
.
A
n
o
ve
r
vi
e
w
of
Y
O
L
O
v8
m
od
e
l
T
h
e
Y
O
L
O
s
e
t
of
m
od
e
l
s
h
a
s
b
e
c
o
m
e
w
e
ll
-
k
no
w
n i
n t
h
e
f
i
e
l
d o
f
c
o
m
pu
te
r
v
i
s
i
o
n.
W
it
h t
h
e
i
n
tr
o
d
uc
ti
o
n
o
f
Y
O
L
O
v
8,
th
e
m
o
s
t
r
e
c
e
nt
v
e
r
s
io
n
of
Y
O
L
O
,
a
m
o
d
e
l
t
ha
t
e
s
t
a
b
li
s
h
e
s
a
n
e
w
s
t
a
t
e
-
of
-
t
h
e
-
a
r
t
f
or
ob
j
e
c
t
r
e
c
og
ni
ti
on
a
n
d
i
n
s
t
a
nc
e
s
e
g
m
e
n
t
a
t
io
n,
t
h
e
f
i
e
l
d
of
c
o
m
pu
te
r
vi
s
i
o
n
a
dv
a
n
c
e
s
.
A
c
c
or
d
i
ng
t
o
c
om
m
o
n
o
bj
e
c
t
s
i
n
c
o
n
te
xt
(
C
O
C
O
)
a
nd
R
o
bo
f
l
o
w
10
0
m
e
a
s
ur
e
m
e
n
t
s
,
Y
O
L
O
v8
h
a
s
a
h
ig
h
a
c
c
ur
a
c
y
r
a
t
e
.
A
v
a
r
ie
t
y
of
d
e
v
e
lo
p
e
r
-
f
r
i
e
n
d
ly
f
e
a
t
ur
e
s
a
r
e
i
n
c
l
u
de
d
i
n
Y
O
L
O
v8
.
O
n
C
O
C
O
,
Y
O
L
O
v
8
a
c
h
i
e
v
e
s
hi
gh
a
c
c
u
r
a
c
y.
F
o
r
i
n
s
t
a
n
c
e
,
w
h
e
n
m
e
a
s
ur
e
d
on
C
O
C
O
,
th
e
m
e
di
um
Y
O
L
O
v
8
m
o
d
e
l
o
bt
a
in
s
a
5
0.
2%
m
e
a
n
a
v
e
r
a
g
e
p
r
e
c
i
s
i
on
(
m
A
P
)
[
23
]
.
D
a
r
k
N
et
-
53,
a
n
e
w
b
a
c
kb
on
e
ne
twor
k
in
tr
od
uc
e
d
in
Y
O
L
O
v
8,
is
s
ub
s
ta
nt
i
a
ll
y
qui
c
k
e
r
a
n
d
m
or
e
pr
e
c
i
s
e
th
a
n
th
e
on
e
u
ti
l
iz
e
d
i
n
Y
O
L
O
v7
.
A
C
N
N
w
it
h
53
l
a
y
e
r
s
c
a
ll
e
d
D
a
r
k
N
e
t
-
53
c
a
n
c
l
a
s
s
if
y
p
ho
to
s
in
to
100
0
di
f
f
e
r
e
n
t
it
e
m
c
a
te
gor
ie
s
.
Y
o
lo
v
8
g
e
n
e
r
a
t
e
s
bo
un
di
n
g
bo
x
pr
e
di
c
ti
on
s
i
n
a
m
a
nn
e
r
a
ki
n
t
o
pi
xe
l
-
w
i
s
e
pi
c
tu
r
e
s
e
g
m
e
nt
a
ti
o
n.
A
ddi
ti
o
n
a
ll
y
e
m
p
lo
yi
n
g
f
e
a
tu
r
e
pyr
a
m
id
n
e
tw
or
k
s
, Y
O
L
O
v8
i
s
m
or
e
a
c
c
ur
a
t
e
ov
e
r
a
ll
a
nd
is
a
bl
e
t
o d
i
s
ti
n
gu
is
h
obj
e
c
t
s
of
v
a
r
i
ou
s
s
iz
e
s
[
24]
.
T
h
e
a
r
c
hi
te
c
t
u
r
e
of
Y
O
L
O
v8
m
o
de
l
i
s
l
i
ke
t
h
e
F
ig
ur
e
2
[
25
]
.
F
ig
ur
e
2
.
A
r
c
hi
te
c
tu
r
e
of
Y
O
L
O
v8 mode
l
4.
R
E
S
U
L
T
S
A
N
D
D
I
S
C
U
S
S
I
O
N
4
.1.
O
u
t
c
om
e
of
t
h
is
r
e
s
e
ar
c
h
T
he
Y
O
L
O
v8
m
od
e
l
ha
s
s
uc
c
e
s
s
f
ul
ly
de
te
c
te
d
te
n
ty
pe
s
of
r
oof
to
ps
w
he
r
e
s
ol
a
r
pa
n
e
ls
c
a
n
be
in
s
ta
ll
e
d. T
he
out
c
om
e
a
f
te
r
ut
il
iz
in
g Y
O
L
O
v8 on the
i
nput
i
m
a
ge
s
i
s
s
how
n i
n
F
ig
ur
e
3.
4
.2.
P
e
r
f
or
m
an
c
e
an
al
ys
is
i
n
t
e
r
m
s
o
f
ac
c
u
r
a
c
y, p
r
e
c
is
io
n
, r
e
c
al
l
an
d
F
1 s
c
or
e
T
hr
e
e
s
e
t
s
-
tr
a
in
in
g,
va
li
da
ti
on,
a
nd
te
s
ti
ng
-
w
e
r
e
c
r
e
a
te
d
f
r
om
t
he
da
ta
s
e
t.
T
he
de
e
p
le
a
r
ni
ng
m
ode
l
is
tr
a
in
e
d
on
tr
a
in
in
g
da
ta
,
va
li
da
te
d
on
va
li
da
ti
on
da
ta
to
c
ha
nge
th
e
hype
r
pa
r
a
m
e
te
r
s
a
nd
m
oni
to
r
th
e
m
ode
l'
s
pe
r
f
or
m
a
nc
e
,
a
nd
te
s
te
d
on
te
s
ti
ng
da
ta
to
de
te
r
m
in
e
how
w
e
ll
th
e
f
in
a
l
m
ode
l
pe
r
f
or
m
s
on
unt
r
ie
d
da
ta
.
80%
of
th
e
im
a
ge
s
w
e
r
e
ut
il
iz
e
d
f
or
tr
a
in
in
g
w
hi
le
th
e
r
e
s
t
of
th
e
im
a
ge
s
w
e
r
e
f
or
va
li
da
ti
on
a
nd
te
s
ti
ng
pur
pos
e
s
.
U
s
in
g
a
te
s
t
d
a
ta
s
e
t
w
it
h
560
da
ta
poi
nt
s
,
th
is
pr
oj
e
c
t'
s
a
c
c
ur
a
c
y
w
a
s
e
xa
m
in
e
d.
U
s
in
g
th
e
pr
e
-
tr
a
in
e
d
m
ode
l
on
ou
r
da
ta
s
e
t,
w
e
w
e
r
e
a
bl
e
to
de
te
c
t
r
oof
to
p
s
w
it
h
a
n
a
c
c
ur
a
c
y
of
93.6%
.
T
he
pr
e
c
is
io
n
-
c
onf
id
e
nc
e
c
ur
ve
pl
ot
s
th
e
pr
e
c
i
s
io
n
of
th
e
m
ode
l'
s
pr
e
di
c
ti
ons
a
ga
in
s
t
di
f
f
e
r
e
nt
c
onf
id
e
nc
e
s
c
or
e
th
r
e
s
hol
ds
.
T
he
pr
e
c
is
io
n c
onf
id
e
nc
e
c
ur
ve
f
or
t
hi
s
r
oof
to
p i
s
s
how
n i
n F
ig
ur
e
4.
T
he
pr
e
c
is
io
n
-
r
e
c
a
ll
c
ur
ve
is
u
s
e
f
ul
f
or
unde
r
s
ta
ndi
ng
how
d
if
f
e
r
e
nt
c
onf
id
e
nc
e
s
c
or
e
th
r
e
s
hol
ds
im
pa
c
t
th
e
pe
r
f
or
m
a
nc
e
of
th
e
obj
e
c
t
de
te
c
ti
on
m
ode
l.
I
t
h
e
l
ps
in
c
hoo
s
in
g
a
n
a
ppr
opr
ia
te
th
r
e
s
hol
d
th
a
t
ba
la
nc
e
s
pr
e
c
is
io
n
a
nd
r
e
c
a
ll
ba
s
e
d
on
th
e
s
pe
c
if
ic
r
e
qui
r
e
m
e
nt
s
of
th
e
a
ppl
ic
a
ti
on.
G
e
ne
r
a
ll
y,
m
ode
l
s
w
it
h
hi
ghe
r
a
r
e
a
s
unde
r
th
e
pr
e
c
is
io
n
-
r
e
c
a
ll
c
ur
ve
a
r
e
c
ons
id
e
r
e
d
be
tt
e
r
pe
r
f
or
m
e
r
s
a
s
th
e
y
c
a
n
a
c
hi
e
ve
hi
gh
pr
e
c
is
io
n w
hi
le
m
a
in
ta
in
in
g good r
e
c
a
ll
. P
r
e
c
is
io
n
-
r
e
c
a
ll
c
ur
ve
is
s
how
n i
n F
ig
ur
e
5.
T
he
F
1
-
s
c
or
e
c
onf
id
e
nc
e
c
ur
ve
is
s
how
n
in
F
ig
ur
e
6.
T
he
F
1
-
s
c
or
e
,
w
hi
c
h
r
a
nge
s
f
r
om
0
to
1,
is
pa
r
ti
c
ul
a
r
ly
us
e
f
ul
in
id
e
nt
if
y
in
g
th
e
le
ve
l
of
c
onf
id
e
nc
e
th
a
t
be
s
t
ba
la
nc
e
s
th
e
pr
e
c
is
io
n
a
nd
r
e
c
a
ll
va
lu
e
s
f
or
a
gi
ve
n
m
ode
l.
T
he
s
e
t
of
F
1
-
s
c
or
e
s
f
or
a
pa
r
ti
c
ul
a
r
m
ode
l
c
a
n
be
us
e
d
to
pr
oduc
e
a
s
in
gl
e
v
a
lu
e
a
s
s
e
s
s
m
e
nt
m
e
a
s
ur
e
, w
hi
c
h m
a
y be
a
r
e
li
a
bl
e
ga
uge
of
t
he
pe
r
f
or
m
a
nc
e
of
t
he
m
ode
l
a
s
a
w
hol
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
3
,
J
une
20
25
:
2282
-
2290
2286
F
ig
ur
e
3
.
R
oof
to
p’
s
d
e
te
c
ti
on uti
li
z
in
g Y
O
L
O
v8 a
lg
or
it
hm
F
ig
ur
e
4
.
P
r
e
c
is
io
n
c
onf
id
e
nc
e
c
ur
ve
F
ig
ur
e
5
.
P
r
e
c
is
io
n
r
e
c
a
ll
c
ur
ve
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
R
oof
to
ps
de
te
c
ti
on w
it
h Y
O
L
O
v
8 f
r
om
ae
r
ia
l
image
r
y
and a br
i
e
f
r
e
v
ie
w
on r
oof
to
p …
(
M
d. Sabbir
A
hm
e
d)
2287
F
ig
ur
e
6
.
F
1 s
c
or
e
c
onf
id
e
nc
e
c
ur
ve
5.
A
S
H
O
R
T
R
E
V
I
E
W ON ROOF
T
O
P
P
H
O
T
O
V
A
L
T
I
C
P
O
T
E
N
T
I
A
L
A
S
S
E
S
S
M
E
N
T
F
ol
lo
w
in
g
th
e
id
e
nt
if
ic
a
ti
on
of
r
oof
to
ps
f
or
th
e
in
s
t
a
ll
a
ti
on
of
P
V
m
odul
e
s
,
th
e
r
e
a
r
e
num
e
r
ous
m
or
e
pos
s
ib
il
it
ie
s
,
in
c
lu
di
ng
m
e
a
s
ur
in
g
th
e
r
oof
to
p'
s
a
r
e
a
,
pow
e
r
ge
ne
r
a
ti
on
c
a
pa
bi
li
ti
e
s
a
nd
de
te
r
m
in
in
g
th
e
num
be
r
of
P
V
m
odul
e
s
th
a
t
c
a
n
be
in
s
ta
ll
e
d
th
e
r
e
.
O
nc
e
m
o
r
e
,
e
s
ti
m
a
ti
ng
s
ol
a
r
ir
r
a
di
a
nc
e
i
s
a
n
e
s
s
e
nt
ia
l
pha
s
e
in
de
te
r
m
in
in
g
how
e
ne
r
gy
-
e
f
f
ic
ie
nt
P
V
m
odul
e
s
a
r
e
. M
a
ny
s
tu
di
e
s
h
a
ve
a
lr
e
a
dy
b
e
e
n
c
onduc
te
d
to
do
s
o.
T
hi
s
s
e
c
ti
on
w
il
l
pr
ovi
de
in
f
or
m
a
ti
on
on
s
tu
di
e
s
th
a
t
ha
ve
di
s
c
ove
r
e
d
pot
e
nt
ia
li
ti
e
s
th
a
t
c
a
n
be
u
s
e
d
a
f
te
r
r
oof
to
p de
te
c
ti
on t
o e
ns
ur
e
pr
ope
r
ut
il
iz
a
ti
on of
s
ol
a
r
e
ne
r
gy.
5.1.
Q
u
an
t
if
yi
n
g s
u
it
ab
le
r
oof
t
op
ar
e
a f
or
as
s
e
s
s
in
g t
h
e
p
ot
e
n
t
ia
li
t
ie
s
of
s
ol
ar
p
ow
e
r
ge
n
e
r
at
io
n
I
t
is
ne
c
e
s
s
a
r
y
to
de
duc
t
th
e
a
r
e
a
oc
c
upi
e
d
by
b
a
r
r
ie
r
s
f
r
om
th
e
to
ta
l
a
r
e
a
w
he
n
c
a
l
c
ul
a
ti
ng
a
r
oof
to
p'
s
e
f
f
e
c
ti
ve
a
r
e
a
.
T
he
r
e
f
or
e
,
a
n
e
f
f
e
c
ti
ve
a
r
e
a
c
a
n
be
m
a
th
e
m
a
ti
c
a
ll
y
de
te
r
m
in
e
d
a
s
th
e
di
f
f
e
r
e
nc
e
be
twe
e
n
th
e
e
nt
ir
e
a
r
e
a
a
nd
th
e
obs
ta
c
le
a
r
e
a
in
pi
xe
ls
,
w
hi
c
h
s
houl
d
be
th
e
n
tr
a
ns
la
te
d
in
to
s
qua
r
e
m
e
te
r
s
.
T
he
ne
c
e
s
s
a
r
y
num
be
r
of
P
V
m
odul
e
s
f
or
a
s
pe
c
if
ic
r
oof
to
p
c
a
n
th
e
n
be
de
te
r
m
in
e
d
by
di
vi
di
ng
th
e
r
oof
to
p'
s
ove
r
a
ll
a
r
e
a
by
th
e
a
r
e
a
of
th
e
P
V
pa
ne
l
th
a
t
w
il
l
be
in
s
ta
ll
e
d
th
e
r
e
.
A
s
ig
ni
f
ic
a
nt
num
be
r
of
s
tu
di
e
s
ha
ve
r
e
ve
a
le
d t
he
r
oof
to
p'
s
pot
e
nt
ia
l
f
or
e
le
c
tr
ic
it
y ge
ne
r
a
ti
on. S
om
e
of
t
he
s
e
s
tu
di
e
s
a
r
e
s
um
m
a
r
iz
e
d
in
T
a
bl
e
1
.
T
a
bl
e
1.
S
um
m
a
r
y of
r
oof
to
p e
le
c
tr
ic
it
y ge
ne
r
a
ti
on pote
nt
ia
l
f
r
o
m
pr
e
vi
ous
s
tu
di
e
s
R
e
f
O
bj
e
c
t
i
ve
U
t
i
l
i
z
e
d
m
e
t
hod
F
i
ndi
ngs
[
18]
Q
ua
nt
i
f
i
c
a
t
i
on of
t
he
s
ui
t
a
bl
e
r
oof
t
op a
r
e
a
.
C
onvol
ut
i
ona
l
ne
ur
a
l
ne
t
w
or
k.
S
e
gm
e
nt
s
ui
t
a
bl
e
r
oof
t
op
a
r
e
a
s
w
i
t
h
a
n
a
c
c
ur
a
c
y
of
93.0%
.
[
24]
E
s
t
i
m
a
t
i
ng
t
he
s
pa
t
i
a
l
di
s
t
r
i
but
i
on
of
s
ol
a
r
PV
pow
e
r
ge
ne
r
a
t
i
on pot
e
nt
i
a
l
De
e
p
l
e
a
r
ni
ng
U
t
i
l
i
z
e
d
U
-
ne
t
m
ode
l
got
a
n
a
c
c
ur
a
c
y of
92%
[
25]
P
r
ovi
di
ng e
s
t
i
m
a
t
i
on of
r
oof
t
op P
V
pow
e
r
ge
ne
r
a
t
i
on
D
e
e
p l
e
a
r
ni
ng
T
he
f
i
ndi
ngs
i
ndi
c
a
t
e
t
ha
t
t
he
pr
ovi
nc
e
'
s
r
oof
t
op
r
e
s
our
c
e
s
ha
ve
a
pot
e
nt
i
a
l
i
ns
t
a
l
l
e
d
c
a
pa
c
i
t
y of
245.17 G
W
.
[
26]
E
s
t
i
m
a
t
i
ng
ut
i
l
i
z
a
bl
e
a
r
e
a
s
a
nd
s
ol
a
r
e
ne
r
gy
pot
e
nt
i
a
l
of
r
oof
t
ops
D
e
e
p
l
e
a
r
ni
ng,
m
or
phol
ogi
c
a
l
ope
r
a
t
i
on, s
e
gm
e
nt
a
t
i
on
A
c
c
ur
a
c
y
:
93%
f
or
r
oof
t
op
e
xt
r
a
c
t
i
on
a
nd
99%
f
o
r
pl
a
ne
s
e
gm
e
nt
a
t
i
on
5.2.
S
ol
ar
ir
r
a
d
ia
n
c
e
e
s
t
im
at
io
n
f
or
t
h
e
as
s
e
s
s
m
e
n
t
of
e
n
e
r
g
y e
f
f
ic
ie
n
c
y o
f
P
V
p
ow
e
r
p
la
n
t
s
b
as
e
d
on
r
oof
s
h
ap
e
an
d
an
gl
e
T
he
a
m
ount
of
e
le
c
tr
om
a
gne
ti
c
r
a
di
a
ti
on
th
a
t
is
e
m
it
te
d
f
r
om
th
e
s
un
pe
r
uni
t
a
r
e
a
,
w
hi
c
h
is
ty
pi
c
a
ll
y
s
qua
r
e
m
e
te
r
s
,
is
known
a
s
s
ol
a
r
ir
r
a
di
a
ti
on.
T
he
s
un
ir
r
a
di
a
ti
on
a
nd
c
e
ll
te
m
pe
r
a
tu
r
e
ha
ve
th
e
bi
gge
s
t
e
f
f
e
c
ts
on
how
m
uc
h
e
l
e
c
tr
ic
it
y
PV
s
our
c
e
s
c
a
n
ge
n
e
r
a
te
.
H
e
r
e
a
r
e
s
om
e
r
e
s
e
a
r
c
he
s
w
hi
c
h
ha
v
e
s
tu
di
e
d on thi
s
t
opi
c
in
T
a
bl
e
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
3
,
J
une
20
25
:
2282
-
2290
2288
T
a
bl
e
2.
S
ol
a
r
i
r
r
a
di
a
nc
e
e
s
ti
m
a
ti
on f
or
r
oof
to
p f
r
om
pr
e
vi
ous
s
tu
di
e
s
Re
f
O
bj
e
c
t
i
v
e
U
t
i
l
i
z
e
d
m
e
t
hod
F
i
ndi
ngs
[27]
F
i
ndi
ng
out
t
he
opt
i
m
u
m
t
i
l
t
a
ngl
e
s
of
m
a
xi
m
u
m
s
ol
a
r
i
rra
di
a
n
c
e
M
a
t
he
m
a
t
i
c
a
l
c
a
l
c
ul
a
t
i
on
T
he
m
a
xi
m
u
m
s
ol
a
r
ra
di
a
t
i
on
c
a
n
be
found
w
i
t
h
t
he
t
i
l
t
a
ngl
e
be
t
w
e
e
n
0°
t
o
64°
[28]
F
i
ndi
ng
out
t
he
e
ffe
c
t
of
t
i
l
t
a
ngl
e
a
nd
ori
e
nt
a
t
i
on
of
s
ol
a
r
s
urfa
c
e
for
s
ol
a
r
po
w
e
r
ge
ne
ra
t
i
on
S
ol
a
r
pa
ne
l
s
a
nd
s
e
ns
ors
T
he
e
ne
rgy
ha
rve
s
t
i
ng
c
a
pa
c
i
t
y
of
e
a
c
h
s
ol
a
r
pa
ne
l
i
s
s
t
rongl
y
i
nfl
ue
nc
e
d
by
t
he
i
nc
l
i
na
t
i
on
[29]
E
s
t
i
m
a
t
i
on
of
a
nnua
l
s
ol
a
r
i
rra
di
a
t
i
on
M
a
c
hi
ne
l
e
a
rn
i
ng
(ra
ndom
fore
s
t
)
A
c
c
ura
c
y
92%
a
t
e
s
t
i
m
a
t
i
ng
s
ol
a
r
i
rra
di
a
t
i
on
[30]
E
s
t
i
m
a
t
i
ng
s
ol
a
r
i
rra
di
a
nc
e
a
nd
PV
pow
e
r
L
ong
s
hort
-
t
e
rm
m
e
m
ory
a
nd
ga
t
e
d
re
c
urre
n
t
uni
t
W
i
t
h
0.
96%,
m
a
c
hi
ne
l
e
a
rni
ng
m
od
e
l
s
produc
e
d
ve
ry
unbi
a
s
e
d
e
s
t
i
m
a
t
i
ons
.
6.
C
O
N
C
L
U
S
I
O
N
I
n
th
is
r
e
s
e
a
r
c
h
w
or
k,
a
s
y
s
te
m
f
or
de
te
c
ti
ng
r
oof
to
ps
s
ui
ta
bl
e
f
or
in
s
ta
ll
a
ti
on
P
V
m
odul
e
s
in
a
e
r
ia
l
im
a
ge
s
us
in
g
de
e
p
le
a
r
ni
ng
te
c
hni
que
s
is
pr
e
s
e
nt
e
d.
T
he
s
tu
d
y
s
ought
to
a
ddr
e
s
s
th
e
dr
a
w
ba
c
k
s
of
m
a
nua
l
in
s
pe
c
ti
ons
a
nd
tr
a
di
ti
ona
l
m
e
th
ods
by a
ppl
yi
ng
de
e
p
le
a
r
ni
ng a
lg
or
it
hm
s
f
or
a
c
c
ur
a
te
a
nd
s
uc
c
e
s
s
f
ul
r
oof
to
p
id
e
nt
if
ic
a
ti
on.
T
he
m
os
t
r
e
c
e
nt
Y
O
L
O
v8
m
ode
l,
bui
lt
a
nd
us
e
d
to
de
te
c
t
pot
e
nt
ia
l
in
s
ta
ll
a
ti
on
s
it
e
s
,
s
how
e
d
pr
om
is
in
g
r
e
s
ul
ts
in
te
r
m
s
of
a
c
c
ur
a
c
y,
pr
e
c
i
s
io
n,
a
nd
F
1
-
s
c
or
e
w
he
n
tr
a
in
e
d
on
a
c
a
r
e
f
ul
ly
s
e
le
c
t
e
d
c
ol
le
c
ti
on
of
a
e
r
ia
l
phot
ogr
a
phs
w
it
h
a
nnot
a
te
d
r
oof
to
p
s
ur
f
a
c
e
s
.
T
h
e
s
ys
t
e
m
'
s
a
ut
om
a
te
d
de
t
e
c
ti
on
pr
oc
e
s
s
,
w
hi
c
h
r
e
duc
e
s
th
e
ti
m
e
a
nd
m
on
e
y
a
s
s
oc
i
a
te
d
w
it
h
m
a
nua
l
in
s
pe
c
ti
ons
,
c
a
n
e
na
bl
e
th
e
f
a
s
t
e
r
in
te
gr
a
ti
on
of
s
ol
a
r
pow
e
r
.
O
nc
e
m
or
e
,
a
pot
e
nt
ia
li
ty
a
na
ly
s
i
s
of
th
is
s
tu
dy
r
e
ve
a
le
d
th
e
e
nor
m
ous
pot
e
nt
ia
l
f
or
s
ol
a
r
e
ne
r
gy
us
e
.
T
he
pol
ic
ym
a
k
e
r
s
,
e
n
e
r
gy
pl
a
nne
r
s
,
a
nd
bui
ld
in
g
ow
ne
r
s
c
a
n
le
ve
r
a
ge
it
s
s
im
pl
ic
it
y
in
e
ne
r
gy
pl
a
nni
ng
a
nd de
c
is
io
n
-
m
a
ki
ng pr
oc
e
s
s
e
s
t
o m
a
k
e
t
he
be
s
t
u
s
e
of
s
ol
a
r
e
n
e
r
gy, whic
h i
s
a
c
r
uc
ia
l
ne
e
d f
or
t
hi
s
c
e
nt
ur
y.
F
U
N
D
I
N
G
I
N
F
O
R
M
A
T
I
O
N
T
he
a
ut
hor
s
de
c
la
r
e
th
a
t
th
is
r
e
s
e
a
r
c
h
w
a
s
c
onduc
te
d
w
it
hout
a
ny
f
in
a
nc
ia
l
s
uppor
t
f
r
om
f
undi
ng
a
ge
nc
ie
s
i
n t
he
publi
c
,
c
om
m
e
r
c
ia
l,
or
not
-
f
or
-
pr
o
f
it
s
e
c
to
r
s
.
A
U
T
H
O
R
C
O
N
T
R
I
B
U
T
I
O
N
S
S
T
A
T
E
M
E
N
T
T
hi
s
jo
ur
na
l
us
e
s
th
e
C
ont
r
ib
ut
or
R
ol
e
s
T
a
xonomy
(
C
R
e
di
T
)
to
r
e
c
ogni
z
e
in
di
vi
dua
l
a
ut
hor
c
ont
r
ib
ut
io
ns
, r
e
duc
e
a
ut
hor
s
hi
p di
s
put
e
s
,
a
nd f
a
c
il
it
a
te
c
ol
la
bo
r
a
ti
on.
N
am
e
o
f
A
u
t
h
or
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
M
d. S
a
bbi
r
A
hm
e
d
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
M
d. S
hohe
l
A
r
m
a
n
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
N
us
r
a
t
T
a
s
ni
m
✓
✓
✓
✓
✓
✓
✓
✓
M
d H
a
f
iz
ul
I
m
r
a
n
✓
✓
✓
✓
✓
✓
✓
M
us
a
bbi
r
H
a
s
a
n S
a
m
m
a
k
✓
✓
✓
✓
✓
✓
T
ouhi
d B
hui
ya
n
✓
✓
✓
✓
✓
✓
✓
✓
C
:
C
onc
e
pt
ua
l
i
z
a
t
i
on
M
:
M
e
t
hodol
ogy
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
da
t
i
on
Fo
:
Fo
r
m
a
l
a
na
l
ys
i
s
I
:
I
nve
s
t
i
ga
t
i
on
R
:
R
e
s
our
c
e
s
D
:
D
a
t
a
C
ur
a
t
i
on
O
:
W
r
i
t
i
ng
-
O
r
i
gi
na
l
D
r
a
f
t
E
:
W
r
i
t
i
ng
-
R
e
vi
e
w
&
E
di
t
i
ng
Vi
:
Vi
s
ua
l
i
z
a
t
i
on
Su
:
Su
pe
r
vi
s
i
on
P
:
P
r
oj
e
c
t
a
dm
i
ni
s
t
r
a
t
i
on
Fu
:
Fu
ndi
ng a
c
qui
s
i
t
i
on
C
O
N
F
L
I
C
T
O
F
I
N
T
E
R
E
S
T
S
T
A
T
E
M
E
N
T
T
he
a
ut
hor
s
de
c
la
r
e
no c
onf
li
c
t
of
i
nt
e
r
e
s
t.
D
A
T
A
A
V
A
I
L
A
B
I
L
I
T
Y
T
he
da
ta
t
ha
t
s
uppor
t
th
e
f
in
di
ngs
of
t
hi
s
s
tu
dy a
r
e
a
va
il
a
bl
e
on
r
e
que
s
t
f
r
om
t
he
f
i
r
s
t
a
ut
hor
, [
M
S
A
]
.
T
he
da
ta
,
w
hi
c
h
c
ont
a
in
in
f
or
m
a
ti
on
th
a
t
c
oul
d
c
om
pr
om
is
e
th
e
pr
iv
a
c
y
of
r
e
s
e
a
r
c
h
p
a
r
ti
c
ip
a
nt
s
,
a
r
e
not
publ
ic
ly
a
va
il
a
bl
e
due
t
o c
e
r
ta
in
r
e
s
tr
ic
ti
ons
.
R
E
F
E
R
E
N
C
E
S
[
1]
A
.
A
boa
h,
B
.
W
a
ng,
U
.
B
a
gc
i
,
a
nd
Y
.
A
du
-
G
ya
m
f
i
,
“
R
e
a
l
-
t
i
m
e
m
ul
t
i
-
c
l
a
s
s
he
l
m
e
t
vi
ol
a
t
i
on
de
t
e
c
t
i
on
us
i
ng
f
e
w
-
s
hot
da
t
a
s
a
m
pl
i
ng
t
e
c
hni
que
a
nd
Y
O
L
O
v8,”
i
n
2023
I
E
E
E
/
C
V
F
C
onf
e
r
e
nc
e
on
C
om
put
e
r
V
i
s
i
on
and
P
at
t
e
r
n
R
e
c
ogni
t
i
on
W
or
k
s
hop
s
(
C
V
P
R
W
)
, I
E
E
E
, J
un. 2023, pp. 5350
–
5358
, doi
:
10.1109/
C
V
P
R
W
59228.2023.00564.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
A
r
ti
f
I
nt
e
ll
I
S
S
N
:
2252
-
8938
R
oof
to
ps
de
te
c
ti
on w
it
h Y
O
L
O
v
8 f
r
om
ae
r
ia
l
image
r
y
and a br
i
e
f
r
e
v
ie
w
on r
oof
to
p …
(
M
d. Sabbir
A
hm
e
d)
2289
[
2]
A
.
D
um
i
t
r
i
u,
F
.
T
a
t
ui
,
F
.
M
i
r
on,
R
.
T
.
I
one
s
c
u,
a
nd
R
.
T
i
m
of
t
e
,
“
R
i
p
c
ur
r
e
nt
s
e
gm
e
nt
a
t
i
on:
a
nove
l
be
nc
hm
a
r
k
a
nd
Y
O
L
O
v8
ba
s
e
l
i
ne
r
e
s
ul
t
s
,”
i
n
2023
I
E
E
E
/
C
V
F
C
onf
e
r
e
nc
e
on
C
om
put
e
r
V
i
s
i
on
and
P
at
t
e
r
n
R
e
c
ogni
t
i
on
W
or
k
s
hops
(
C
V
P
R
W
)
,
I
E
E
E
,
J
un. 2023, pp. 1261
–
1271
, doi
:
10.1109/
C
V
P
R
W
59228.2023.00133.
[
3]
J.
-
H
. K
i
m
, N
. K
i
m
, a
nd C
. S
. W
on, “
H
i
gh
-
s
pe
e
d dr
one
d
e
t
e
c
t
i
on ba
s
e
d on
Y
O
L
O
-
v8,”
i
n
I
C
A
SSP
2023
-
2023 I
E
E
E
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on
A
c
ou
s
t
i
c
s
,
Spe
e
c
h
and
Si
gnal
P
r
oc
e
s
s
i
ng
(
I
C
A
SSP
)
,
I
E
E
E
,
J
un.
2023,
pp.
1
–
2
,
doi
:
10.1109/
I
C
A
S
S
P
49357.2023.10095516.
[
4]
Y
.
L
i
,
Q
.
F
a
n,
H
.
H
ua
ng, Z
.
H
a
n,
a
nd
Q
.
G
u,
“
A
m
odi
f
i
e
d
Y
O
L
O
v8
de
t
e
c
t
i
on
ne
t
w
or
k
f
or
ua
v
a
e
r
i
a
l
i
m
a
ge
r
e
c
ogni
t
i
on,”
D
r
one
s
,
vol
. 7, no. 5, M
a
y 2023, doi
:
10.3390/
dr
one
s
7050304.
[
5]
N
.
B
uhl
,
“
Y
O
L
O
obj
e
c
t
de
t
e
c
t
i
on
e
xpl
a
i
ne
d:
e
vol
ut
i
on,
a
l
gor
i
t
hm
,
a
nd
a
p
pl
i
c
a
t
i
ons
,
”
e
N
C
O
R
D
.
A
c
c
e
s
s
e
d:
A
ug.
25,
2023.
[
O
nl
i
ne
]
. A
va
i
l
a
bl
e
:
ht
t
ps
:
/
/
e
nc
or
d.c
om
/
bl
og/
yol
o
-
obj
e
c
t
-
de
t
e
c
t
i
on
-
gui
de
[
6]
T
.
Z
hong
e
t
al
.
,
“
A
c
i
t
y
-
s
c
a
l
e
e
s
t
i
m
a
t
i
on
of
r
oof
t
op
s
ol
a
r
phot
ovol
t
a
i
c
pot
e
nt
i
a
l
ba
s
e
d
on
de
e
p
l
e
a
r
ni
ng,”
A
ppl
i
e
d
E
ne
r
gy
,
vol
. 298, S
e
p. 2021, doi
:
10.1016/
j
.a
pe
ne
r
gy.2021.117132.
[
7]
H
.
M
a
o
e
t
al
.
,
“
A
dva
nc
e
s
a
nd
pr
os
pe
c
t
s
on
e
s
t
i
m
a
t
i
ng
s
ol
a
r
phot
ovol
t
a
i
c
i
n
s
t
a
l
l
a
t
i
on
c
a
pa
c
i
t
y
a
nd
pot
e
nt
i
a
l
ba
s
e
d
on
s
a
t
e
l
l
i
t
e
a
nd
a
e
r
i
a
l
i
m
a
ge
s
,”
R
e
ne
w
abl
e
and Su
s
t
ai
nabl
e
E
ne
r
gy
R
e
v
i
e
w
s
, vol
. 179, J
un. 2023, doi
:
10.1016/
j
.r
s
e
r
.2023.113276.
[
8]
M
.
A
s
l
a
ni
a
nd
S
.
S
e
i
pe
l
,
“
A
ut
om
a
t
i
c
i
de
nt
i
f
i
c
a
t
i
on
of
ut
i
l
i
z
a
bl
e
r
oof
t
op
a
r
e
a
s
i
n
di
gi
t
a
l
s
ur
f
a
c
e
m
ode
l
s
f
or
phot
ovol
t
a
i
c
s
po
t
e
nt
i
a
l
a
s
s
e
s
s
m
e
nt
,”
A
ppl
i
e
d E
ne
r
gy
, vol
. 306, J
a
n. 2022, doi
:
10.1016/
j
.a
pe
ne
r
gy.202
1.118033.
[
9]
N
.
M
oha
j
e
r
i
,
D
.
A
s
s
oul
i
ne
,
B
.
G
ui
boud,
A
.
B
i
l
l
,
A
.
G
udm
unds
s
on,
a
nd
J
.
-
L
.
S
c
a
r
t
e
z
z
i
ni
,
“
A
c
i
t
y
-
s
c
a
l
e
r
oof
s
ha
pe
c
l
a
s
s
i
f
i
c
a
t
i
on
us
i
ng
m
a
c
hi
ne
l
e
a
r
ni
ng
f
or
s
ol
a
r
e
ne
r
gy
a
ppl
i
c
a
t
i
ons
,”
R
e
ne
w
abl
e
E
ne
r
gy
,
vol
.
121,
pp.
81
–
93,
J
un.
2018,
doi
:
10.1016/
j
.r
e
ne
ne
.2017.12.096.
[
10]
X
.
S
ong
e
t
al
.
,
“
A
n
a
ppr
oa
c
h
f
or
e
s
t
i
m
a
t
i
ng
s
ol
a
r
phot
ovol
t
a
i
c
pot
e
nt
i
a
l
ba
s
e
d
on
r
oof
t
op
r
e
t
r
i
e
va
l
f
r
om
r
e
m
ot
e
s
e
ns
i
ng
i
m
a
ge
s
,
”
E
ne
r
gi
e
s
, vol
. 11, no. 11, N
ov. 2018, doi
:
10.3390/
e
n11113172.
[
11]
S
.
L
e
e
,
S
.
I
ye
nga
r
,
M
.
F
e
ng,
P
.
S
he
noy,
a
nd
S
.
M
a
j
i
,
“
D
e
e
p
R
oof
:
a
da
t
a
-
dr
i
ve
n
a
ppr
oa
c
h
f
or
s
ol
a
r
pot
e
nt
i
a
l
e
s
t
i
m
a
t
i
on
u
s
i
ng
r
oof
t
op
i
m
a
ge
r
y,”
i
n
P
r
oc
e
e
di
ngs
of
t
he
25t
h
A
C
M
SI
G
K
D
D
I
nt
e
r
nat
i
onal
C
o
nf
e
r
e
nc
e
on
K
now
l
e
dge
D
i
s
c
ov
e
r
y
&
D
at
a
M
i
ni
ng
,
N
e
w
Y
or
k, U
S
A
:
A
C
M
, J
ul
. 2019, pp. 2105
–
2113
, doi
:
10.1145/
3292500.3330741.
[
12]
T
.
Z
hong
e
t
al
.
,
“
A
s
s
e
s
s
m
e
nt
of
s
ol
a
r
phot
ovol
t
a
i
c
pot
e
nt
i
a
l
s
on
ur
ba
n
noi
s
e
ba
r
r
i
e
r
s
us
i
ng
s
t
r
e
e
t
-
vi
e
w
i
m
a
ge
r
y,”
R
e
ne
w
abl
e
E
ne
r
gy
, vol
. 168, pp. 181
–
194, M
a
y 2021, doi
:
10.1016/
j
.r
e
ne
ne
.2020.12.044.
[
13]
S
.
P
r
a
bha
ka
r
a
n,
R
.
A
nni
e
U
t
hr
a
,
a
nd
J
.
P
r
e
e
t
ha
r
os
e
l
yn,
“
D
e
e
p
l
e
a
r
ni
ng
-
ba
s
e
d
m
ode
l
f
or
de
f
e
c
t
de
t
e
c
t
i
on
a
nd
l
oc
a
l
i
z
a
t
i
on
o
n
phot
ovol
t
a
i
c
pa
ne
l
s
,”
C
om
put
e
r
Sy
s
t
e
m
s
Sc
i
e
nc
e
and
E
ngi
ne
e
r
i
ng
,
vol
.
44,
no.
3,
pp.
2683
–
2700,
2023,
doi
:
10.32604/
c
s
s
e
.2023.028898.
[
14]
A
.
M
e
m
a
r
i
,
V
.
C
.
D
a
m
,
a
nd
L
.
N
ol
l
e
,
“
D
e
e
p
l
e
a
r
ni
ng
f
or
de
t
e
c
t
i
ng
t
i
l
t
a
ngl
e
a
nd
or
i
e
nt
a
t
i
on
of
phot
ovol
t
a
i
c
pa
ne
l
s
on
s
a
t
e
l
l
i
t
e
i
m
a
ge
r
y,”
i
n
A
r
t
i
f
i
c
i
al
I
nt
e
l
l
i
ge
nc
e
X
X
X
I
X
:
42nd
SG
A
I
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on
A
r
t
i
f
i
c
i
al
I
nt
e
l
l
i
ge
nc
e
,
2022,
pp.
255
–
266
,
doi
:
10.1007/
978
-
3
-
031
-
21441
-
7_18.
[
15]
D
.
C
he
n,
J
.
X
i
a
o,
a
nd
N
.
L
i
,
“
A
l
ow
-
c
o
s
t
de
e
p
l
e
a
r
ni
ng
f
r
a
m
e
w
or
k
f
or
phot
ov
ol
t
a
i
c
pow
e
r
pl
a
nt
de
t
e
c
t
i
on
i
n
r
e
a
l
w
or
l
d,”
Soc
i
al
Sc
i
e
nc
e
R
e
s
e
a
r
c
h N
e
t
w
or
k
(
SSR
N
)
,
pp. 1
-
20,
2023
, doi
:
10.2139/
s
s
r
n.4496797.
[
16]
H
.
T
e
l
l
a
,
M
.
M
oha
nde
s
,
B
.
L
i
u,
S
.
R
e
hm
a
n,
a
nd
A
.
A
l
-
S
ha
i
khi
,
“
D
e
e
p
l
e
a
r
n
i
ng
s
ys
t
e
m
f
or
de
f
e
c
t
c
l
a
s
s
i
f
i
c
a
t
i
on
of
s
ol
a
r
pa
ne
l
c
e
l
l
s
,”
i
n
2022
14t
h
I
nt
e
r
nat
i
onal
C
onf
e
r
e
n
c
e
on
C
om
put
at
i
onal
I
nt
e
l
l
i
ge
n
c
e
and
C
o
m
m
uni
c
at
i
on
N
e
t
w
or
k
s
(
C
I
C
N
)
,
I
E
E
E
,
D
e
c
. 2022, pp. 448
–
453
, doi
:
10.1109/
C
I
C
N
56167.2022.10008277.
[
17]
E
.
F
a
khr
a
i
a
n,
M
.
A
.
F
or
m
e
nt
,
F
.
V
.
D
a
l
m
a
u,
A
.
N
a
m
e
ni
,
a
nd
M
.
J
.
C
.
G
ue
r
r
e
r
o
,
“
D
e
t
e
r
m
i
na
t
i
on
of
t
he
ur
ba
n
r
oof
t
op
phot
ovol
t
a
i
c
pot
e
nt
i
a
l
:
a
s
t
a
t
e
of
t
he
a
r
t
,”
E
ne
r
gy
R
e
por
t
s
, vol
. 7, pp. 176
–
185, S
e
p. 2021, do
i
:
10.1016/
j
.e
gyr
.2021.06.031.
[
18]
R
.
C
a
s
t
e
l
l
o,
A
.
W
a
l
c
h,
R
.
A
t
t
i
a
s
,
R
.
C
a
d
e
i
,
S
.
J
i
a
ng,
a
nd
J
.
-
L
.
S
c
a
r
t
e
z
z
i
ni
,
“
Q
ua
nt
i
f
i
c
a
t
i
on
of
t
he
s
ui
t
a
bl
e
r
oof
t
op
a
r
e
a
f
or
s
ol
a
r
pa
ne
l
i
ns
t
a
l
l
a
t
i
on f
r
om
ove
r
he
a
d i
m
a
ge
r
y u
s
i
ng c
onvol
ut
i
ona
l
ne
ur
a
l
n
e
t
w
or
ks
,
”
J
our
nal
of
P
hy
s
i
c
s
:
C
onf
e
r
e
n
c
e
Se
r
i
e
s
, vol
.
2042,
no. 1, N
ov. 2021, doi
:
10.1088/
1742
-
6596/
2042/
1/
012002.
[
19]
A
.
K
.
B
e
hur
a
,
A
.
K
um
a
r
,
D
.
K
.
R
a
j
a
k,
C
.
I
.
P
r
unc
u,
a
nd
L
.
L
a
m
be
r
t
i
,
“
T
ow
a
r
ds
be
t
t
e
r
pe
r
f
or
m
a
nc
e
s
f
or
a
nove
l
r
oof
t
op
s
ol
a
r
PV
s
ys
t
e
m
,”
Sol
ar
E
ne
r
gy
, vol
. 216, pp. 518
–
529, M
a
r
. 2021, doi
:
10.1016/
j
.s
ol
e
ne
r
.2021.01.045.
[
20]
R
.
C
a
d
e
i
,
R
.
A
t
t
i
a
s
,
a
nd
S
.
J
i
a
ng,
“
D
e
t
e
c
t
i
ng
r
oof
t
op
a
va
i
l
a
bl
e
s
ur
f
a
c
e
f
or
i
ns
t
a
l
l
i
ng
P
V
m
odul
e
s
i
n
a
e
r
i
a
l
i
m
a
ge
s
us
i
ng
de
e
p
l
e
a
r
ni
ng,”
E
nv
i
r
onm
e
nt
al
Sc
i
e
nc
e
J
ou
r
nal
,
pp. 1
-
5,
2020.
[
21]
L
.
K
.
W
i
gi
nt
on,
H
.
T
.
N
guye
n,
a
nd
J
.
M
.
P
e
a
r
c
e
,
“
Q
ua
nt
i
f
yi
ng
r
oof
t
op
s
ol
a
r
phot
ovol
t
a
i
c
pot
e
nt
i
a
l
f
or
r
e
gi
ona
l
r
e
ne
w
a
bl
e
e
ne
r
gy
pol
i
c
y,”
C
om
put
e
r
s
,
E
nv
i
r
onm
e
nt
and
U
r
ban
Sy
s
t
e
m
s
,
vol
.
34,
no.
4,
pp.
345
–
357,
J
ul
.
2010,
doi
:
10.1016/
j
.c
om
pe
nvur
bs
ys
.2010.01.001.
[
22]
A
.
A
.
A
.
G
a
s
s
a
r
a
nd
S
.
H
.
C
ha
,
“
R
e
vi
e
w
of
ge
ogr
a
phi
c
i
nf
or
m
a
t
i
on
s
ys
t
e
m
s
-
ba
s
e
d
r
oof
t
op
s
ol
a
r
phot
ovol
t
a
i
c
pot
e
nt
i
a
l
e
s
t
i
m
a
t
i
on
a
ppr
oa
c
he
s
a
t
ur
ba
n s
c
a
l
e
s
,”
A
ppl
i
e
d E
ne
r
gy
, vol
. 291, J
un. 2021, doi
:
10.1016/
j
.a
pe
ne
r
gy.2021.116817.
[
23]
H
.
L
ou
e
t
al
.
,
“
D
C
-
Y
O
L
O
V
8:
s
m
a
l
l
-
s
i
z
e
obj
e
c
t
de
t
e
c
t
i
on
a
l
gor
i
t
hm
ba
s
e
d
on
c
a
m
e
r
a
s
e
ns
or
,”
E
l
e
c
t
r
oni
c
s
,
vol
.
12,
no.
10
,
M
a
y 2023, doi
:
10.3390/
e
l
e
c
t
r
oni
c
s
12102323.
[
24]
T
.
S
un,
M
.
S
ha
n,
X
.
R
ong,
a
nd
X
.
Y
a
ng,
“
E
s
t
i
m
a
t
i
ng
t
he
s
pa
t
i
a
l
di
s
t
r
i
but
i
on
o
f
s
ol
a
r
phot
ovol
t
a
i
c
pow
e
r
ge
ne
r
a
t
i
on
pot
e
nt
i
a
l
on
di
f
f
e
r
e
nt
t
ype
s
of
r
ur
a
l
r
oof
t
ops
us
i
ng
a
de
e
p
l
e
a
r
ni
ng
ne
t
w
or
k
a
ppl
i
e
d
t
o
s
a
t
e
l
l
i
t
e
i
m
a
ge
s
,”
A
ppl
i
e
d
E
ne
r
g
y
,
vol
.
315,
2022
,
doi
:
10.1016/
j
.a
pe
ne
r
gy.2022.119025.
[
25]
H
.
J
i
a
ng
e
t
al
.
,
“
G
e
os
pa
t
i
a
l
a
s
s
e
s
s
m
e
nt
of
r
oof
t
op
s
ol
a
r
phot
ovol
t
a
i
c
pot
e
nt
i
a
l
us
i
ng
m
ul
t
i
-
s
our
c
e
r
e
m
ot
e
s
e
ns
i
ng
d
a
t
a
,”
E
ne
r
g
y
and A
I
, vol
. 10, 2022, doi
:
10.1016/
j
.e
gya
i
.2022.100185.
[
26]
M
. A
s
l
a
ni
a
nd S
.
S
e
i
pe
l
, “
A
s
pa
t
i
a
l
l
y d
e
t
a
i
l
e
d a
ppr
oa
c
h
t
o t
he
a
s
s
e
s
s
m
e
nt
of
r
o
of
t
op s
ol
a
r
e
ne
r
gy pot
e
nt
i
a
l
b
a
s
e
d
on l
i
da
r
da
t
a
,”
i
n
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on
G
e
ogr
aphi
c
al
I
nf
or
m
at
i
on
Sy
s
t
e
m
s
T
he
or
y
,
A
ppl
i
c
at
i
ons
and
M
anage
m
e
nt
,
G
I
ST
A
M
,
2022,
pp. 56
–
63
, doi
:
10.5220/
0011108300003185.
[
27]
Q
.
H
a
s
s
a
n,
M
.
K
.
A
bba
s
,
A
.
M
.
A
bdul
a
t
e
e
f
,
J
.
A
bul
a
t
e
e
f
,
a
nd
A
.
M
oha
m
a
d
,
“
A
s
s
e
s
s
m
e
nt
t
he
pot
e
nt
i
a
l
s
ol
a
r
e
n
e
r
gy
w
i
t
h
t
h
e
m
ode
l
s
f
or
opt
i
m
um
t
i
l
t
a
ngl
e
s
of
m
a
xi
m
um
s
ol
a
r
i
r
r
a
di
a
nc
e
f
or
I
r
a
q
,”
C
as
e
St
udi
e
s
i
n
C
he
m
i
c
al
and
E
nv
i
r
onm
e
nt
al
E
ngi
ne
e
r
i
ng
,
vol
. 4, 2021, doi
:
10.1016/
j
.c
s
c
e
e
.2021.100140.
[
28]
E
.
Y
ul
i
z
a
,
L
.
L
i
z
a
l
i
di
a
w
a
t
i
,
a
nd
R
.
E
ka
w
i
t
a
,
“
T
he
e
f
f
e
c
t
of
t
i
l
t
a
ngl
e
a
nd
o
r
i
e
nt
a
t
i
on
of
s
ol
a
r
s
ur
f
a
c
e
on
s
ol
a
r
r
oof
t
op
m
i
ni
a
t
ur
e
s
ys
t
e
m
i
n
be
ngkul
u
uni
ve
r
s
i
t
y,”
I
nt
e
r
nat
i
onal
J
our
nal
of
E
ne
r
gy
and
E
nv
i
r
onm
e
nt
al
E
ngi
ne
e
r
i
ng
,
vol
.
12,
no.
3,
pp.
589
–
598,
2021, doi
:
10.1007/
s
40095
-
021
-
00390
-
4.
[
29]
A
.
W
a
l
c
h,
R
.
C
a
s
t
e
l
l
o,
N
.
M
oh
a
j
e
r
i
,
a
nd
J
.
-
L
.
S
c
a
r
t
e
z
z
i
ni
,
“
A
f
a
s
t
m
a
c
hi
ne
l
e
a
r
ni
ng
m
ode
l
f
or
l
a
r
ge
-
s
c
a
l
e
e
s
t
i
m
a
t
i
on
of
a
nnua
l
s
ol
a
r
i
r
r
a
di
a
t
i
on
on
r
oof
t
ops
,”
i
n
P
r
oc
e
e
di
ngs
of
t
he
I
S
E
S
Sol
ar
W
or
l
d
C
ongr
e
s
s
2019
,
F
r
e
i
bur
g,
G
e
r
m
a
ny:
I
nt
e
r
na
t
i
ona
l
S
ol
a
r
E
ne
r
gy S
oc
i
e
t
y, 2019, pp. 1
–
10
, doi
:
10.18086/
s
w
c
.2019.45.12.
[
30]
R
.
A
.
A
.
R
a
m
a
dha
n,
Y
.
R
.
J
.
H
e
a
t
ubun,
S
.
F
.
T
a
n,
a
nd
H
.
J
.
L
e
e
,
“
C
om
pa
r
i
s
on
of
phys
i
c
a
l
a
nd
m
a
c
hi
ne
l
e
a
r
ni
ng
m
ode
l
s
f
or
e
s
t
i
m
a
t
i
ng
s
ol
a
r
i
r
r
a
di
a
nc
e
a
nd
phot
ovol
t
a
i
c
pow
e
r
,”
R
e
ne
w
abl
e
E
ne
r
gy
,
vol
.
178,
pp.
1006
–
1019,
2021,
doi
:
10.1016/
j
.r
e
ne
ne
.2021.06.079.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
A
r
ti
f
I
nt
e
ll
,
V
ol
.
14
, N
o.
3
,
J
une
20
25
:
2282
-
2290
2290
B
I
O
G
R
A
P
H
I
E
S
O
F
A
U
T
H
O
R
S
Md.
Sabbir
Ahmed
has
completed
his
B.Sc
.
from
Departm
ent
of
Software
Engineering,
Daffodil
International
University.
He
is
a
dedicated
professional
in
the
field
of
data
science.
With
a
passion
for
leveraging
data,
his
specializes
in
a
wid
e
range
of
skills,
includin
g
Python
programming,
data
visualization
using
tableau
,
and
advanced
t
echniques
in
deep
learning
and
machine
learning.
His
academic
and
practical
experience
make
him
a
valuable
asset
in
the
world
of
data
science,
where
he
continually
seeks
innovative
solution
s
to
real
-
world
challenges.
He can be contacted at email:
sabbir35
-
2820@
diu.edu.bd.
Md.
Shohel
Arman
has
completed
his
B.Sc.
and
M.Sc.
fro
m
Department
of
Softwa
re
Engine
ering,
Daff
odil
Inter
nation
Univer
sity.
He
is
curre
nt
ly
workin
g
as
an
assista
nt
professor
in
the
Department
of
Software
Engineering,
Daffodil
Inter
national
University.
He
is
also
lab
in
charge
of
the
DIU
Data
Science
Lab.
He
is
currently
worki
ng
with
analyzing
data
for
business
objectives,
data
visualization
(Google
Data
Studio/
Microso
ft
Power
BI,
and
Tableau
,
Python
),
image
processing,
and
deep
learning
.
He
can
be
contacted
at
email:
arman.swe@
diu.edu.
bd.
Nusrat
Tasnim
has
completed
her
B.Sc.
and
M.Sc.
from
Institut
e
of
Information
Technology
,
Jahangirnagar
University.
She
is
currently
working
as
a
lecturer
in
the
Department
of
Software
Engineering,
Daffodil
International
University.
Her
rese
arch
interests
are
machine
learning,
deep
learning
,
NLP
,
and
computer
vision
.
She
can
be
contacted
at
email:
nusrattasnim17@
gmail.com
.
Md
Hafizul
Imran
obtained
his
bachelor
degree
in
electrica
l
and
electronic
engineerin
g
from
Daffodil
International
University,
master’s
degree
in
computer
science
from
Jahangirnaga
r
University,
Banglade
sh
.
Currently
he
is
working
as
a
senior
lecturer
at
Daffodil
International
University
.
Now
he
pursuing
his
Ph
.
D
.
d
egree
in
roboti
cs
field
at
the
Universiti
of
Sains
Malaysia
(USM),
Engineering
Campus,
Nibong
Tebal,
Pena
ng,
Malaysia.
He
can
be
contacted
at email
: hafizul
.swe@
daffodilv
arsity.ed
u.bd
.
Musabb
ir
Hasan
Sammak
has
completed
his
B.Sc.
from
M
awlana
Bhashani
Scienc
e
Techn
ology
Univer
sity.
He
has
comple
ted
in
M.Sc.
in
data
science
and
anal
ytics
fro
m
Universiti
Sains
Malaysia.
He
is
currently
working
as
a
lecturer
in
t
he
Department
of
Software
Engineering,
Daffodil
International
University.
His
research
interests
are
data
analytics,
machine
learning,
deep lea
rning
, and NLP. He can be contac
ted at email: musabbir.swe@
diu.edu.bd
.
Touhid
Bhuiyan
is
an
experienced
Professor
with
a
demons
trated
history
of
working
in
the
education
management
industry.
Skilled
in
cyber
security,
database,
big
data
analytics
,
analytical
skills
,
lecturin
g
.
Strong
engineering
professional
with
a
Certificate
focused
in
Cyber
Security
from
University
of
Oxford.
He
can
be
contacted
at
email:
touhid.bhuiyan@wust.edu
.
Evaluation Warning : The document was created with Spire.PDF for Python.