I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
15
,
No
.
2
,
A
p
r
il
2
0
2
6
,
p
p
.
1
4
2
8
~
1
4
4
0
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
2
.
p
p
1
4
2
8
-
1
4
4
0
1428
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Dro
ne
-
a
ss
isted
de
ep learning
w
eed
detec
tion for sus
t
a
ina
ble
a
g
riculture
a
nd
e
nv
iro
nmenta
l resi
lience
Ag
us
t
a
n L
a
t
if
1,
2
,
H
a
nd
a
ru
J
a
t
i
3
,
H
er
ma
n Dw
i Surj
o
no
3
,
M
a
ni Yu
s
uf
4
1
D
o
c
t
o
r
a
l
P
r
o
g
r
a
m
o
f
E
n
g
i
n
e
e
r
i
n
g
S
c
i
e
n
c
e
s,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
si
t
a
s
N
e
g
e
r
i
Y
o
g
y
a
k
a
r
t
a
,
Y
o
g
y
a
k
a
r
t
a
,
I
n
d
o
n
e
si
a
2
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
S
y
s
t
e
ms
,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
si
t
a
s
M
u
s
a
m
u
s,
M
e
r
a
u
k
e
,
I
n
d
o
n
e
si
a
3
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s E
n
g
i
n
e
e
r
i
n
g
a
n
d
I
n
f
o
r
mat
i
c
s
E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
E
n
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
si
t
a
s Ne
g
e
r
i
Y
o
g
y
a
k
a
r
t
a
,
Y
o
g
y
a
k
a
r
t
a
,
I
n
d
o
n
e
si
a
4
D
e
p
a
r
t
me
n
t
o
f
A
g
r
o
t
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
a
s
M
u
s
a
mu
s
,
M
e
r
a
u
k
e
,
I
n
d
o
n
e
si
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
27
,
2
0
2
5
R
ev
is
ed
J
an
9
,
2
0
2
6
Acc
ep
ted
J
an
25
,
2
0
2
6
Eff
e
c
ti
v
e
we
e
d
d
e
tec
ti
o
n
p
lay
s
a
c
ru
c
ial
ro
le
in
su
sta
in
a
b
le
a
g
ricu
lt
u
re
,
b
o
o
st
in
g
c
ro
p
p
ro
d
u
c
ti
v
it
y
a
n
d
s
u
p
p
o
rti
n
g
e
n
v
ir
o
n
m
e
n
tal
c
o
n
se
rv
a
ti
o
n
.
T
h
is
stu
d
y
c
o
m
p
a
re
s
th
re
e
d
e
e
p
lea
rn
in
g
m
o
d
e
ls
—
YO
LOv
5
,
YO
LO
-
NA
S
,
a
n
d
m
a
sk
re
g
io
n
-
b
a
se
d
c
o
n
v
o
lu
t
io
n
a
l
n
e
u
ra
l
n
e
two
r
k
(
M
a
sk
R
-
CNN
)
-
a
g
a
in
st
trad
it
io
n
a
l
m
e
th
o
d
s
in
term
s
o
f
a
c
c
u
ra
c
y
,
p
ro
c
e
ss
in
g
sp
e
e
d
,
a
n
d
a
d
a
p
tab
il
it
y
in
tr
o
p
ica
l
a
g
ric
u
lt
u
ra
l
c
o
n
d
it
i
o
n
s,
with
M
e
ra
u
k
e
,
I
n
d
o
n
e
sia
,
a
s
th
e
c
a
se
stu
d
y
.
Th
e
r
e
su
lt
s
s
h
o
w
t
h
a
t
Y
OLO
-
NA
S
d
e
li
v
e
rs
t
h
e
h
ig
h
e
st
a
c
c
u
ra
c
y
a
t
9
6
%
with
a
p
r
o
c
e
ss
in
g
t
ime
o
f
2
5
m
s
p
e
r
ima
g
e
,
m
a
k
in
g
it
su
it
a
b
le
fo
r
h
ig
h
-
p
re
c
isio
n
a
p
p
li
c
a
ti
o
n
s.
YO
LOv
5
b
a
lan
c
e
s
stro
n
g
a
c
c
u
ra
c
y
(9
4
%
)
with
fa
ste
r
p
ro
c
e
ss
in
g
a
t
1
2
m
s
p
e
r
ima
g
e
,
e
sta
b
li
sh
in
g
i
t
a
s
t
h
e
m
o
st
e
ffe
c
ti
v
e
fo
r
re
a
l
-
ti
m
e
sc
e
n
a
rio
s.
M
a
sk
R
-
CNN
a
lso
a
c
h
iev
e
s
9
4
%
a
c
c
u
ra
c
y
a
n
d
p
r
o
v
i
d
e
s
a
d
v
a
n
c
e
d
se
g
m
e
n
tatio
n
c
a
p
a
b
il
i
ti
e
s,
b
u
t
it
s slo
we
r
p
ro
c
e
ss
in
g
s
p
e
e
d
o
f
3
1
m
s
li
m
it
s
larg
e
-
sc
a
le
imp
lem
e
n
tatio
n
.
Tra
d
it
i
o
n
a
l
m
e
th
o
d
s
p
e
rf
o
rm
p
o
o
rl
y
i
n
c
o
m
p
a
riso
n
,
wit
h
o
n
ly
8
5
%
a
c
c
u
ra
c
y
a
n
d
p
ro
c
e
ss
in
g
ti
m
e
a
b
o
v
e
5
0
m
s
p
e
r
ima
g
e
.
Th
e
se
fin
d
i
n
g
s
h
i
g
h
li
g
h
t
th
e
tran
sfo
rm
a
ti
v
e
p
o
ten
ti
a
l
o
f
a
rti
ficia
l
in
telli
g
e
n
c
e
(
AI
)
-
b
a
se
d
we
e
d
d
e
t
e
c
ti
o
n
f
o
r
p
re
c
isio
n
a
g
ric
u
lt
u
re
,
p
a
rti
c
u
larly
in
tr
o
p
ica
l
re
g
io
n
s
l
ik
e
M
e
ra
u
k
e
.
Ad
o
p
ti
o
n
o
f
m
o
d
e
ls
s
u
c
h
a
s
YO
LOv
5
re
d
u
c
e
s
m
a
n
u
a
l
lab
o
r
d
e
p
e
n
d
e
n
c
e
wh
il
e
a
d
v
a
n
c
in
g
e
fficie
n
t,
e
c
o
-
frien
d
ly
we
e
d
m
a
n
a
g
e
m
e
n
t.
F
u
tu
re
re
se
a
rc
h
sh
o
u
ld
e
x
p
a
n
d
d
a
tas
e
ts
a
n
d
e
x
p
lo
re
n
e
we
r
m
o
d
e
ls
li
k
e
YO
LOv
8
,
Y
OLO
-
NA
S
,
v
isi
o
n
tran
sf
o
rm
e
rs
(
ViTs)
,
a
n
d
h
y
b
rid
a
p
p
r
o
a
c
h
e
s.
K
ey
w
o
r
d
s
:
C
o
m
p
u
ter
v
is
io
n
Dee
p
lear
n
in
g
Pre
cisi
o
n
ag
r
icu
ltu
r
e
W
ee
d
d
etec
tio
n
YOL
O
-
NAS
YOL
Ov
5
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Ag
u
s
tan
L
atif
Do
cto
r
al
Pro
g
r
a
m
o
f
E
n
g
in
ee
r
in
g
Scien
ce
,
Facu
lty
o
f
E
n
g
in
ee
r
in
g
,
Un
iv
e
r
s
itas
Neg
er
i Y
o
g
y
ak
ar
ta
Yo
g
y
ak
ar
ta,
I
n
d
o
n
es
ia
E
m
ail: a
g
u
s
tan
@
u
n
m
u
s
.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
Ag
r
icu
ltu
r
e
p
la
y
s
a
v
ital
r
o
le
i
n
en
s
u
r
in
g
en
v
ir
o
n
m
e
n
tal
s
u
s
tain
ab
ilit
y
an
d
g
lo
b
al
f
o
o
d
s
ec
u
r
ity
.
T
h
is
s
ec
to
r
is
th
e
m
ain
f
o
cu
s
in
I
n
d
o
n
esia,
esp
ec
ially
in
th
e
g
o
v
er
n
m
en
t'
s
ef
f
o
r
ts
to
r
ea
lize
M
er
au
k
e
as
a
n
atio
n
al
f
o
o
d
b
ar
n
.
Me
r
au
k
e
h
as
ex
ce
llen
t
p
o
ten
tial
as
a
f
o
o
d
p
r
o
d
u
ctio
n
ce
n
ter
t
h
an
k
s
to
its
v
a
s
t
f
er
tile
lan
d
an
d
s
tr
ateg
ic
lo
ca
tio
n
f
o
r
ag
r
icu
lt
u
r
al
d
e
v
elo
p
m
e
n
t.
Ho
wev
e
r
,
th
e
p
r
esen
ce
o
f
wee
d
s
is
a
s
er
io
u
s
o
b
s
tacle
th
at
r
ed
u
ce
s
cr
o
p
p
r
o
d
u
ctiv
ity
[
1
]
,
[
2
]
.
W
ee
d
s
n
o
t
o
n
l
y
co
m
p
ete
with
th
e
m
ain
cr
o
p
in
ter
m
s
o
f
n
u
tr
ie
n
ts
,
wate
r
,
an
d
lig
h
t
b
u
t
also
in
cr
e
ase
p
r
o
d
u
ctio
n
c
o
s
ts
th
r
o
u
g
h
th
e
n
ee
d
f
o
r
s
p
r
ay
in
g
h
er
b
icid
es
an
d
ad
d
itio
n
al
lab
o
r
f
o
r
m
an
u
al
co
n
tr
o
l
[
3
]
–
[
5
]
.
E
f
f
e
c
t
i
v
e
h
e
r
b
i
c
i
d
e
u
s
e
i
s
o
n
e
o
f
t
h
e
k
e
y
f
a
c
t
o
r
s
f
o
r
m
a
n
a
g
i
n
g
we
e
d
s
i
n
a
g
r
i
c
u
l
t
u
r
al
f
i
e
l
d
s
[
6
]
.
O
n
t
h
e
o
t
h
e
r
h
a
n
d
,
u
s
i
n
g
h
e
r
b
i
c
i
d
e
s
s
i
g
n
i
f
i
ca
n
t
l
y
n
e
g
a
t
i
v
e
l
y
i
m
p
a
c
ts
t
h
e
e
n
v
i
r
o
n
m
e
n
t
[
7
]
.
T
h
e
r
e
f
o
r
e
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dro
n
e
-
a
s
s
is
ted
d
ee
p
lea
r
n
in
g
w
ee
d
d
etec
tio
n
fo
r
s
u
s
ta
in
a
b
le
a
g
r
icu
ltu
r
e
a
n
d
…
(
A
g
u
s
ta
n
La
tif
)
1429
i
n
n
o
v
a
t
i
v
e
a
p
p
r
o
a
c
h
e
s
a
r
e
n
e
e
d
e
d
t
o
a
d
d
r
e
s
s
t
h
is
i
s
s
u
e
e
f
f
e
c
t
i
v
e
l
y
a
n
d
e
f
f
i
ci
e
n
t
l
y
w
h
i
l
e
e
n
s
u
r
i
n
g
t
h
e
s
u
s
ta
i
n
a
b
i
li
t
y
o
f
a
g
r
i
c
u
l
t
u
r
al
e
co
s
y
s
t
e
m
s
[
8
]
.
I
n
r
ec
en
t
y
ea
r
s
,
ar
tific
ial
in
tellig
en
ce
(
AI
)
tech
n
o
lo
g
y
an
d
d
e
ep
lear
n
in
g
h
a
v
e
d
ev
elo
p
ed
r
a
p
id
ly
an
d
o
f
f
er
p
o
te
n
tial
s
o
lu
tio
n
s
f
o
r
f
aster
an
d
m
o
r
e
ac
cu
r
ate
we
ed
d
etec
tio
n
[
9
]
–
[
1
1
]
.
B
ey
o
n
d
ef
f
icien
c
y
,
th
ese
m
eth
o
d
s
ca
n
r
ed
u
ce
ch
em
ic
al
u
s
e
an
d
m
an
u
al
la
b
o
r
,
b
u
t
p
r
ac
tical
d
e
p
lo
y
m
e
n
t
s
till
r
eq
u
ir
es
b
alan
cin
g
ac
cu
r
ac
y
an
d
s
p
ee
d
[
1
2
]
,
[
1
3
]
.
T
h
er
ef
o
r
e,
alg
o
r
ith
m
s
ar
e
alwa
y
s
ev
alu
ated
o
n
th
e
i
m
ag
e
lev
el
o
f
b
o
th
ac
cu
r
ac
y
an
d
s
p
ee
d
[
1
4
]
–
[
2
1
]
.
I
n
teg
r
atin
g
wee
d
d
etec
tio
n
tech
n
o
lo
g
y
i
n
f
o
o
d
cr
o
p
d
ev
elo
p
m
en
t
ca
n
b
e
a
s
tr
ateg
ic
s
tep
to
in
cr
ea
s
e
ag
r
ic
u
ltu
r
al
p
r
o
d
u
ctiv
ity
i
n
r
ea
lizin
g
th
e
v
is
io
n
o
f
s
u
s
tain
ab
ilit
y
.
W
e
e
d
c
o
n
t
r
o
l
is
a
g
l
o
b
a
l
c
h
a
ll
en
g
e
i
n
a
g
r
i
c
u
l
t
u
r
e
a
n
d
h
a
s
a
v
e
r
y
s
i
g
n
i
f
i
c
a
n
t
i
m
p
ac
t
o
n
c
r
o
p
p
r
o
d
u
c
t
i
o
n
.
G
l
o
b
a
l
l
y
,
w
ee
d
s
a
r
e
r
es
p
o
n
s
ib
l
e
f
o
r
3
4
%
o
f
c
r
o
p
y
i
e
l
d
l
o
s
s
es
,
m
a
k
i
n
g
t
h
e
m
o
n
e
o
f
th
e
m
a
i
n
t
h
r
e
a
ts
t
o
a
g
r
i
c
u
l
t
u
r
a
l
p
r
o
d
u
c
t
i
v
i
t
y
.
I
n
Me
r
a
u
k
e
,
t
h
i
s
p
r
o
b
l
e
m
is
e
x
a
c
e
r
b
a
t
e
d
b
y
t
h
e
v
as
t
s
c
a
le
o
f
f
a
r
m
l
a
n
d
t
h
a
t
i
s
d
i
f
f
i
c
u
lt
t
o
m
o
n
i
t
o
r
m
an
u
ally
,
i
n
c
r
e
a
s
i
n
g
t
h
e
n
e
e
d
f
o
r
e
f
f
i
c
i
e
n
t
a
u
t
o
m
a
t
e
d
s
o
l
u
t
i
o
n
s
.
V
a
r
i
o
u
s
w
e
ed
d
e
t
e
c
t
i
o
n
m
e
t
h
o
d
s
h
a
v
e
b
e
e
n
d
e
v
e
l
o
p
e
d
,
i
n
c
l
u
d
i
n
g
c
o
n
v
e
n
t
i
o
n
a
l
d
i
g
i
t
a
l
i
m
a
g
e
-
b
a
s
e
d
m
e
t
h
o
d
s
[
2
2
]
–
[
2
4
]
,
w
h
i
c
h
a
r
e
o
f
t
e
n
c
o
n
s
t
r
a
i
n
e
d
b
y
l
i
m
it
a
t
i
o
n
s
i
n
p
r
o
c
e
s
s
i
n
g
s
p
ee
d
a
n
d
i
d
e
n
tif
i
c
a
t
i
o
n
ac
c
u
r
a
c
y
[
2
5
]
.
H
o
we
v
e
r
,
a
p
p
l
y
i
n
g
t
h
is
t
e
c
h
n
o
l
o
g
y
i
n
t
r
o
p
i
c
a
l
a
g
r
i
c
u
l
t
u
r
a
l
e
n
v
i
r
o
n
m
e
n
t
s
s
u
c
h
a
s
M
e
r
a
u
k
e
o
f
t
e
n
f
a
c
e
s
c
h
a
l
l
e
n
g
e
s
,
i
n
c
l
u
d
i
n
g
w
e
e
d
h
e
t
e
r
o
g
e
n
e
i
t
y
,
y
i
e
l
d
l
o
s
s
e
s
,
a
n
d
w
e
e
d
m
a
n
a
g
e
m
e
n
t
s
t
r
a
t
e
g
ies
t
h
a
t
r
el
y
h
e
a
v
i
l
y
o
n
b
r
o
a
d
h
e
r
b
i
c
i
d
e
a
p
p
l
i
ca
t
i
o
n
c
a
n
i
n
t
e
n
s
i
f
y
e
n
v
i
r
o
n
m
e
n
t
a
l
p
r
e
s
s
u
r
e
o
n
s
o
i
l
a
n
d
w
at
e
r
s
y
s
t
em
s
,
a
n
d
d
y
n
a
m
i
c
l
i
g
h
t
i
n
g
c
o
n
d
i
t
i
o
n
s
,
w
h
i
c
h
a
f
f
e
ct
d
e
t
e
c
t
i
o
n
s
p
e
e
d
a
n
d
a
c
c
u
r
a
c
y
.
T
h
e
r
e
f
o
r
e
,
p
r
a
c
t
i
c
a
l
w
e
e
d
d
e
t
ec
t
i
o
n
t
o
o
l
s
m
u
s
t
n
o
t
o
n
l
y
b
e
a
c
c
u
r
a
t
e
b
u
t
a
l
s
o
f
a
s
t
e
n
o
u
g
h
t
o
s
u
p
p
o
r
t
t
i
m
e
l
y
,
s
i
t
e
-
s
p
e
c
i
f
i
c
i
n
t
e
r
v
e
n
t
i
o
n
s
i
n
t
h
e
f
iel
d
.
T
h
i
s
r
es
e
a
r
c
h
e
v
al
u
a
t
es
a
n
d
c
o
m
p
a
r
e
s
v
a
r
i
o
u
s
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
s
e
d
w
e
e
d
d
e
t
e
c
ti
o
n
m
e
t
h
o
d
s
,
i
n
c
l
u
d
i
n
g
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
(
C
NN
)
,
Y
O
L
O
v
5
,
Y
O
L
O
-
NA
S
,
a
n
d
m
a
s
k
r
e
g
i
o
n
-
b
a
s
e
d
co
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
(
M
a
s
k
R
-
C
NN
)
,
r
e
g
a
r
d
i
n
g
a
c
c
u
r
a
c
y
a
n
d
d
e
t
e
c
t
i
o
n
s
p
ee
d
.
T
h
e
r
e
s
e
a
r
c
h
f
o
c
u
s
e
s
o
n
a
n
a
l
y
z
i
n
g
t
h
e
a
d
v
a
n
t
a
g
e
s
o
f
t
h
e
s
e
m
e
t
h
o
d
s
o
v
e
r
t
r
a
d
i
ti
o
n
a
l
w
ee
d
d
e
t
e
cti
o
n
t
e
c
h
n
i
q
u
e
s
,
w
h
i
c
h
a
r
e
o
f
t
e
n
c
o
n
s
t
r
ai
n
e
d
b
y
l
i
m
i
t
at
i
o
n
s
i
n
te
r
m
s
o
f
e
f
f
i
c
ie
n
c
y
a
n
d
p
r
e
c
i
s
i
o
n
.
T
a
k
i
n
g
i
n
t
o
a
c
c
o
u
n
t
t
h
e
c
h
a
r
a
c
t
e
r
is
t
ics
o
f
a
g
r
i
c
u
l
t
u
r
e
i
n
M
e
r
a
u
k
e
,
w
h
i
c
h
is
a
t
r
o
p
i
c
a
l
r
eg
i
o
n
,
t
h
i
s
s
t
u
d
y
a
ls
o
a
i
m
s
t
o
i
d
e
n
t
i
f
y
t
h
e
m
o
s
t
a
d
a
p
t
i
v
e
d
e
e
p
le
a
r
n
i
n
g
m
e
t
h
o
d
s
f
o
r
l
o
c
a
l
c
o
n
d
i
ti
o
n
s
.
T
h
e
r
e
s
u
l
ts
o
f
t
h
e
s
t
u
d
y
a
r
e
e
x
p
e
c
t
e
d
t
o
s
i
g
n
if
i
c
a
n
t
l
y
c
o
n
t
r
i
b
u
t
e
t
o
t
h
e
d
e
v
elo
p
m
e
n
t
o
f
A
I
-
b
a
s
e
d
a
g
r
i
c
u
l
t
u
r
a
l
t
e
c
h
n
o
l
o
g
i
es
t
h
at
a
r
e
m
o
r
e
e
f
f
e
ct
i
v
e
,
e
f
f
i
c
i
e
n
t
,
a
n
d
r
e
l
e
v
a
n
t
t
o
s
u
p
p
o
r
t
s
u
s
ta
i
n
a
b
l
e
a
g
r
i
c
u
l
t
u
r
a
l
p
r
o
d
u
c
t
i
v
i
t
y
,
b
o
t
h
i
n
M
e
r
a
u
k
e
a
n
d
g
l
o
b
a
l
l
y
.
I
n
a
d
d
i
t
i
o
n
,
t
h
i
s
r
e
s
e
a
r
c
h
w
i
ll
u
n
d
e
r
s
c
o
r
e
t
h
e
i
m
p
o
r
t
a
n
c
e
o
f
t
e
c
h
n
o
l
o
g
i
c
a
l
s
o
l
u
t
i
o
n
s
t
h
a
t
i
n
cr
e
a
s
e
p
r
o
d
u
c
t
i
v
it
y
a
n
d
s
t
r
e
n
g
t
h
e
n
t
h
e
s
u
s
t
a
i
n
a
b
il
i
t
y
o
f
e
n
v
i
r
o
n
m
e
n
t
a
l
e
c
o
s
y
s
t
e
m
s
.
Alth
o
u
g
h
wee
d
d
etec
tio
n
r
esear
ch
h
as
e
x
p
an
d
ed
,
cr
o
s
s
-
m
eth
o
d
b
en
ch
m
ar
k
in
g
th
at
co
n
tr
asts
s
tate
-
of
-
th
e
-
ar
t
d
ee
p
lear
n
i
n
g
m
o
d
els
with
tr
ad
itio
n
al
tec
h
n
iq
u
es
u
n
d
e
r
f
ield
-
r
ea
lis
tic
tr
o
p
ical
co
n
d
itio
n
s
r
em
ain
s
lim
ited
.
C
o
n
v
en
tio
n
al
ap
p
r
o
ac
h
es
o
f
ten
f
ac
e
s
p
ee
d
an
d
ac
c
u
r
ac
y
lim
its
in
lar
g
e,
h
eter
o
g
en
eo
u
s
f
ield
s
[
2
6
]
,
[
2
7
]
,
wh
er
ea
s
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
es
(
e.
g
.
,
C
NNs
an
d
YOL
Ov
5
)
h
av
e
r
ep
o
r
ted
u
p
to
~9
5
%
ac
cu
r
ac
y
in
s
ev
er
al
s
tu
d
ies
[
2
8
]
.
New
e
r
m
o
d
els
s
u
ch
as
YOL
O
-
NA
S
an
d
Ma
s
k
R
-
C
NN
p
r
o
m
is
e
im
p
r
o
v
ed
f
ea
t
u
r
e
lear
n
in
g
an
d
s
eg
m
en
tatio
n
c
ap
ab
ilit
y
.
Ho
wev
er
,
th
e
y
h
a
v
e
y
et
to
b
e
c
o
m
p
r
e
h
en
s
iv
ely
an
aly
ze
d
in
a
cr
o
s
s
-
m
eth
o
d
co
m
p
ar
is
o
n
b
y
co
n
s
id
er
in
g
tr
o
p
ical
en
v
ir
o
n
m
en
tal
ch
ar
ac
ter
is
tics
an
d
ec
o
s
y
s
tem
s
u
s
tain
ab
ilit
y
[
2
9
]
,
[
3
0
]
.
T
h
er
ef
o
r
e,
th
is
s
tu
d
y
f
ills
th
e
g
ap
b
y
ev
alu
atin
g
th
e
p
er
f
o
r
m
a
n
ce
o
f
c
u
ttin
g
-
ed
g
e
d
ee
p
lea
r
n
in
g
m
eth
o
d
s
ag
ai
n
s
t
tr
ad
itio
n
al
tech
n
iq
u
es,
b
o
th
in
s
p
ee
d
a
n
d
a
cc
u
r
ac
y
,
to
p
r
o
v
id
e
n
ew
in
s
ig
h
ts
f
o
r
d
ev
elo
p
in
g
m
o
r
e
ad
a
p
tiv
e
an
d
s
u
s
tain
ab
le
wee
d
d
etec
tio
n
s
o
lu
tio
n
s
.
T
h
is
s
tu
d
y
m
ak
es
a
m
ajo
r
co
n
tr
ib
u
tio
n
to
ca
r
ef
u
lly
ev
alu
ati
n
g
th
r
ee
d
ee
p
lear
n
in
g
m
eth
o
d
s
,
n
am
ely
YOL
Ov
5
,
YOL
O
-
NAS,
an
d
Ma
s
k
R
-
C
NN
in
wee
d
d
e
tectio
n
wh
ile
co
m
p
ar
in
g
th
e
m
with
tr
ad
itio
n
al
tech
n
iq
u
es.
Pre
v
io
u
s
r
esear
ch
h
as
s
h
o
wn
th
at
d
ee
p
lear
n
in
g
-
b
ased
m
eth
o
d
s
h
av
e
h
i
g
h
ac
c
u
r
ac
y
p
o
te
n
tial
f
o
r
ex
am
p
le,
YOL
Ov
5
r
ec
o
r
d
ed
an
ac
cu
r
ac
y
o
f
o
v
er
9
0
%
i
n
o
b
ject
d
etec
tio
n
[
3
1
]
–
[
3
3
]
.
A
h
an
d
s
-
o
n
s
tu
d
y
co
m
p
ar
in
g
th
e
ad
v
a
n
tag
es
o
f
t
h
is
m
eth
o
d
with
co
n
v
en
tio
n
al
tech
n
iq
u
es
in
ter
m
s
o
f
s
p
ee
d
a
n
d
ac
cu
r
ac
y
is
s
till
m
in
im
al,
esp
ec
ially
f
o
r
r
eg
io
n
s
with
u
n
iq
u
e
ch
allen
g
es
s
u
ch
as
wee
d
h
eter
o
g
e
n
eity
an
d
v
ar
ied
lig
h
tin
g
co
n
d
itio
n
s
[
3
4
]
.
B
y
ad
ap
tin
g
t
h
e
m
o
d
el
to
en
v
ir
o
n
m
e
n
tal
co
n
d
itio
n
s
,
th
e
s
tu
d
y
ex
p
an
d
s
ac
ad
em
ic
in
s
ig
h
ts
i
n
ap
p
ly
in
g
cu
ttin
g
-
e
d
g
e
tech
n
o
l
o
g
ies
in
th
e
ag
r
icu
ltu
r
al
s
ec
t
o
r
[
3
5
]
,
[
3
6
]
.
T
h
is
s
tu
d
y
ev
al
u
ated
a
d
ataset
o
f
d
r
o
n
e
-
b
ased
wee
d
im
ag
er
y
f
r
o
m
Me
r
au
k
e,
I
n
d
o
n
esia
(
1
,
3
1
8
p
r
im
ar
y
im
ag
es),
c
o
m
b
in
e
d
with
an
ad
d
itio
n
al
s
ec
o
n
d
ar
y
im
a
g
e
s
et
(
4
,
0
5
0
im
ag
es)
to
im
p
r
o
v
e
r
esil
ien
ce
u
n
d
er
tr
o
p
ical
f
ield
v
ar
iab
ilit
y
.
T
r
ain
in
g
an
d
em
p
ir
ical
ev
alu
atio
n
o
f
th
e
m
o
d
el,
f
o
llo
wed
b
y
c
o
n
tex
tu
al
co
m
p
ar
is
o
n
s
with
th
e
latest
liter
atu
r
e
to
p
lace
f
in
d
in
g
s
in
b
r
o
ad
er
r
esear
c
h
tr
en
d
s
.
T
h
is
r
esear
ch
p
r
o
v
id
es
r
elev
an
t
p
r
ac
tical
s
o
lu
tio
n
s
to
s
u
p
p
o
r
t
s
u
s
tain
ab
le
ag
r
icu
ltu
r
e
[
3
7
]
.
T
h
e
r
esu
lts
o
f
th
is
s
tu
d
y
a
r
e
ex
p
e
cted
to
an
s
wer
th
e
u
r
g
en
t
n
ee
d
f
o
r
m
o
r
e
ef
f
icien
t
an
d
ac
cu
r
ate
wee
d
d
etec
tio
n
m
eth
o
d
s
in
tr
o
p
ical
r
eg
io
n
s
,
esp
ec
i
ally
Me
r
au
k
e.
2.
M
E
T
H
O
D
T
h
is
s
tu
d
y
ad
o
p
ts
a
h
y
b
r
id
em
p
ir
ical
an
d
co
n
tex
tu
al
b
en
c
h
m
ar
k
in
g
d
esig
n
.
W
e
tr
ain
ed
an
d
ev
alu
ated
YOL
Ov
5
,
YOL
O
-
NAS,
an
d
Ma
s
k
R
-
C
NN
o
n
a
co
m
b
in
ed
wee
d
im
ag
e
d
ataset
co
n
s
is
tin
g
o
f
1
,
3
1
8
p
r
im
ar
y
d
r
o
n
e
im
a
g
es
co
llected
in
M
er
au
k
e,
I
n
d
o
n
esia
an
d
4
,
0
5
0
s
ec
o
n
d
ar
y
im
a
g
es
(
all
at
5
2
8
0
×2
9
7
0
r
eso
lu
tio
n
)
.
Mo
d
el
p
er
f
o
r
m
an
ce
was
ass
e
s
s
ed
u
s
in
g
s
tan
d
ar
d
class
if
ica
tio
n
an
d
d
etec
tio
n
m
etr
ics
as
well
as
in
f
er
en
ce
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
4
2
8
-
1
4
4
0
1430
tim
e,
an
d
co
m
p
a
r
ed
ag
ain
s
t
a
tr
ad
itio
n
al
b
aselin
e.
T
o
p
o
s
itio
n
th
e
ex
p
er
im
en
tal
f
in
d
i
n
g
s
with
in
b
r
o
ad
er
r
esear
ch
tr
en
d
s
,
we
also
r
ev
iewe
d
r
ec
en
t
p
ee
r
-
r
e
v
iewe
d
s
tu
d
ies
r
ep
o
r
tin
g
co
m
p
ar
a
b
le
ac
cu
r
ac
y
an
d
s
p
ee
d
m
etr
ics
f
o
r
wee
d
d
etec
tio
n
.
T
h
is
ap
p
r
o
ac
h
allo
ws
th
e
id
en
tific
atio
n
o
f
th
e
ad
v
a
n
tag
es
an
d
d
is
ad
v
an
tag
es
o
f
ea
ch
m
eth
o
d
b
ased
o
n
ac
cu
r
ac
y
,
s
p
ee
d
,
an
d
ad
ap
tab
ili
ty
to
tr
o
p
ical
en
v
ir
o
n
m
en
ts
s
u
ch
as
Me
r
au
k
e,
th
u
s
m
ak
in
g
a
s
ig
n
if
ican
t
co
n
tr
ib
u
tio
n
to
th
e
o
p
tim
izatio
n
o
f
ag
r
icu
ltu
r
al
tech
n
o
lo
g
y
an
d
im
p
r
o
v
in
g
th
e
d
ec
is
io
n
-
m
ak
in
g
p
r
o
ce
s
s
in
ag
r
icu
ltu
r
e
[
3
8
]
–
[
4
2
]
.
2
.
1
.
Resea
rc
h
d
esig
n
T
h
e
d
esig
n
o
f
th
is
s
tu
d
y
an
aly
ze
d
p
r
im
ar
y
an
d
s
ec
o
n
d
ar
y
d
ata
with
a
q
u
an
titativ
e
a
p
p
r
o
ac
h
to
ev
alu
ate
th
e
p
er
f
o
r
m
an
ce
o
f
v
ar
io
u
s
d
ee
p
lear
n
in
g
-
b
ased
wee
d
d
etec
tio
n
m
eth
o
d
s
.
B
ased
o
n
p
r
ev
io
u
s
liter
atu
r
e
s
tu
d
ies,
tech
n
iq
u
es
s
u
ch
as
C
NN,
YOL
Ov
5
,
YOL
O
-
NAS,
an
d
Ma
s
k
R
-
C
NN
h
av
e
b
ee
n
s
h
o
wn
t
o
b
e
ef
f
ec
tiv
e
in
a
v
a
r
iety
o
f
o
b
j
ec
t
d
etec
tio
n
s
ce
n
ar
io
s
,
in
clu
d
in
g
wee
d
id
en
tific
atio
n
,
b
u
t
w
ith
v
ar
y
in
g
d
eg
r
ee
s
o
f
ac
cu
r
ac
y
a
n
d
s
p
ee
d
[
4
3
]
–
[
4
5
]
.
T
h
e
r
esear
c
h
wo
r
k
f
lo
w
c
o
m
p
r
is
es
f
o
u
r
s
tag
es:
i)
d
r
o
n
e
im
ag
e
ac
q
u
is
itio
n
an
d
cu
r
atio
n
o
f
a
co
m
p
lem
e
n
tar
y
s
ec
o
n
d
ar
y
im
a
g
e
s
et,
ii)
an
n
o
tatio
n
an
d
p
r
ep
r
o
ce
s
s
in
g
,
iii)
m
o
d
el
tr
ain
in
g
an
d
ev
alu
atio
n
o
n
h
eld
-
o
u
t
d
a
ta,
an
d
iv
)
co
n
tex
t
u
al
co
m
p
ar
i
s
o
n
with
r
ec
en
t
liter
atu
r
e.
T
h
e
o
v
er
all
wo
r
k
f
lo
w
is
s
u
m
m
ar
ized
in
Fig
u
r
e
1
.
Fo
r
t
h
e
em
p
i
r
i
ca
l
co
m
p
o
n
en
t,
th
e
co
m
b
i
n
e
d
d
atas
et
w
as
o
r
g
an
i
ze
d
i
n
t
o
tr
ai
n
i
n
g
,
v
al
id
ati
o
n
,
a
n
d
tes
t
s
u
b
s
ets
.
I
m
ag
es
we
r
e
p
r
e
p
r
o
ce
s
s
e
d
t
o
m
a
tc
h
t
h
e
in
p
u
t
r
eq
u
i
r
e
m
e
n
ts
o
f
e
ac
h
m
o
d
el,
an
d
s
t
a
n
d
ar
d
d
a
ta
au
g
m
e
n
ta
ti
o
n
was
ap
p
l
ie
d
t
o
i
m
p
r
o
v
e
r
o
b
u
s
t
n
ess
to
ill
u
m
i
n
a
tio
n
an
d
b
a
c
k
g
r
o
u
n
d
v
a
r
i
ati
o
n
.
Fo
r
t
h
e
co
n
t
e
x
t
u
al
co
m
p
o
n
e
n
t
,
a
s
tr
u
ct
u
r
ed
lit
er
a
tu
r
e
r
e
v
i
ew
w
as
c
o
n
d
u
c
te
d
i
n
r
e
p
u
ta
b
l
e
d
at
ab
ases
(
e
.
g
.
,
I
E
E
E
Xp
l
o
r
e,
E
ls
e
v
i
er
,
an
d
Sp
r
i
n
g
e
r
L
in
k
)
u
s
i
n
g
k
e
y
w
o
r
d
s
s
u
c
h
as
“
we
ed
d
et
ec
t
io
n
”,
“
d
e
ep
l
ea
r
n
in
g
”,
“Y
OL
O
”,
a
n
d
“M
as
k
R
-
C
NN
”
,
an
d
s
tu
d
ies
we
r
e
r
et
ai
n
e
d
wh
e
n
th
e
y
r
e
p
o
r
te
d
q
u
a
n
t
ita
ti
v
e
ac
cu
r
a
cy
a
n
d
s
p
ee
d
m
e
tr
ics
c
o
m
p
a
r
a
b
l
e
to
t
h
is
w
o
r
k
.
Fig
u
r
e
1
.
R
esear
ch
d
esig
n
with
p
r
im
ar
y
an
d
s
ec
o
n
d
a
r
y
d
ata
2
.
2
.
Da
t
a
c
o
llect
io
n a
nd
pro
ce
s
s
ing
Prim
ar
y
d
ata
wer
e
co
llected
in
Me
r
au
k
e,
I
n
d
o
n
esia
u
s
in
g
a
d
r
o
n
e
(
u
n
m
a
n
n
ed
ae
r
ial
v
eh
icle
or
UAV)
.
T
h
e
p
r
im
ar
y
d
ataset
co
n
tain
s
1
,
3
1
8
im
ag
es
with
v
ar
iatio
n
s
in
illu
m
in
atio
n
,
v
i
ewin
g
an
g
le,
an
d
b
ac
k
g
r
o
u
n
d
.
I
n
ad
d
itio
n
,
a
s
e
co
n
d
ar
y
d
ataset
o
f
4
,
0
5
0
im
a
g
es
was
cu
r
ated
to
co
m
p
lem
e
n
t
th
e
p
r
im
ar
y
d
ata
an
d
in
c
r
ea
s
e
v
ar
ia
b
ilit
y
.
All
i
m
ag
es
h
av
e
a
s
p
atial
r
eso
lu
ti
o
n
o
f
5
2
8
0
×2
9
7
0
p
ix
els.
W
e
ed
ca
teg
o
r
ies
wer
e
d
ef
in
ed
b
ased
o
n
h
ab
itat
a
n
d
m
o
r
p
h
o
lo
g
ical
c
h
ar
ac
ter
is
tics
to
r
ef
lect
f
ield
c
o
n
d
itio
n
s
.
An
n
o
tatio
n
s
wer
e
p
r
ep
ar
e
d
f
o
r
m
o
d
el
tr
ain
in
g
a
n
d
ev
alu
atio
n
,
in
clu
d
in
g
b
o
u
n
d
in
g
-
b
o
x
lab
els
f
o
r
YOL
O
-
b
ased
d
etec
to
r
s
an
d
in
s
tan
ce
-
lev
el
lab
els f
o
r
Ma
s
k
R
-
C
NN.
Se
co
n
d
ar
y
d
at
a
wa
s
co
ll
ec
ted
th
r
o
u
g
h
s
y
s
tem
at
ic
l
i
ter
atu
r
e
s
e
ar
ch
e
s
in
r
ep
u
tab
le
d
a
tab
a
s
es
s
u
ch
a
s
I
E
E
E
Xp
lo
r
e,
E
l
s
ev
i
er
,
an
d
Sp
r
in
g
er
L
in
k
.
T
h
e
d
a
ta
s
o
u
g
h
t
in
c
lu
d
e
s
tu
d
i
e
s
r
el
ev
an
t
to
th
e
ap
p
l
ic
at
io
n
o
f
d
ee
p
l
ea
r
n
in
g
f
o
r
we
ed
d
e
t
ec
t
io
n
,
s
p
ec
if
i
ca
lly
th
e
C
N
N
s
,
Y
OL
Ov
5
,
Y
OL
O
-
NA
S,
an
d
Ma
s
k
R
-
C
NN
m
eth
o
d
s
.
T
h
e
s
ea
r
ch
wa
s
c
o
n
d
u
ct
ed
u
s
in
g
s
tr
a
te
g
i
ca
lly
ar
r
an
g
ed
k
ey
wo
r
d
s
,
s
u
ch
a
s
“
w
ee
d
d
e
te
ct
io
n
”,
“
d
ee
p
l
ea
r
n
in
g
m
e
th
o
d
s
”,
“
a
c
cu
r
ac
y
”,
an
d
“
p
r
o
ce
s
s
in
g
s
p
ee
d
”.
T
h
e
r
e
s
u
l
t
s
o
f
th
e
l
i
ter
atu
r
e
s
ea
r
ch
wer
e
th
en
s
e
le
ct
ed
b
a
s
ed
o
n
i
n
cl
u
s
io
n
a
n
d
ex
c
lu
s
io
n
cr
it
er
i
a
to
en
s
u
r
e
th
at
th
e
d
at
a
u
s
ed
wer
e
r
e
le
v
an
t
to
t
h
e
r
es
ea
r
ch
o
b
je
ct
iv
e
s
.
S
tu
d
i
e
s
th
at
m
e
et
t
h
e
cr
it
er
i
a
ar
e
th
o
s
e
th
a
t
p
r
e
s
en
t
q
u
an
t
i
ta
tiv
e
d
a
ta
r
el
at
ed
t
o
ac
cu
r
ac
y
(
%)
an
d
p
r
o
ce
s
s
i
n
g
s
p
e
ed
(
m
s
/f
r
am
e
o
r
FPS
)
,
a
s
r
ec
o
m
m
e
n
d
ed
in
p
r
ev
io
u
s
s
tu
d
ie
s
[
4
6
]
.
Qu
an
t
it
at
i
v
e
li
ter
ac
y
w
as
t
h
e
b
as
i
s
o
f
s
el
ec
tio
n
to
m
a
in
t
ai
n
th
e
ac
cu
r
a
cy
an
d
v
a
l
id
i
ty
o
f
th
e
d
a
ta
an
aly
s
i
s
[
4
7
]
.
A
r
ti
cl
es
th
a
t
d
id
n
o
t
in
c
lu
d
e
p
er
f
o
r
m
an
ce
m
e
tr
ic
s
o
r
o
n
ly
f
o
cu
s
ed
o
n
th
eo
r
e
ti
ca
l
a
s
p
ec
t
s
w
it
h
o
u
t
ex
p
er
i
m
e
n
t
s
w
er
e
ex
clu
d
ed
f
r
o
m
f
u
r
th
e
r
an
aly
s
i
s
.
T
r
ain
in
g
f
o
llo
wed
t
h
e
o
f
f
icial
im
p
lem
en
tatio
n
s
o
f
ea
ch
m
o
d
el.
T
o
e
n
s
u
r
e
f
ai
r
co
m
p
ar
is
o
n
,
th
e
s
am
e
cu
r
ated
d
ataset
an
d
ev
alu
atio
n
p
r
o
to
co
l
wer
e
u
s
ed
ac
r
o
s
s
m
o
d
els.
Mo
d
el
-
s
p
ec
if
ic
co
n
f
i
g
u
r
atio
n
s
(
o
p
tim
izer
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dro
n
e
-
a
s
s
is
ted
d
ee
p
lea
r
n
in
g
w
ee
d
d
etec
tio
n
fo
r
s
u
s
ta
in
a
b
le
a
g
r
icu
ltu
r
e
a
n
d
…
(
A
g
u
s
ta
n
La
tif
)
1431
lear
n
in
g
-
r
ate
s
ch
e
d
u
le,
in
p
u
t
r
esizin
g
,
an
d
au
g
m
e
n
tatio
n
)
f
o
llo
wed
th
e
r
ec
o
m
m
en
d
e
d
b
aselin
e
s
ettin
g
s
p
r
o
v
id
e
d
b
y
ea
c
h
f
r
a
m
ewo
r
k
,
with
m
o
d
el
s
elec
tio
n
b
ased
o
n
v
alid
atio
n
p
e
r
f
o
r
m
an
ce
.
I
n
f
er
en
ce
tim
e
was
m
ea
s
u
r
ed
co
n
s
is
ten
tly
to
r
ep
r
e
s
en
t p
r
ac
tical
d
ep
lo
y
m
en
t c
o
n
s
tr
ain
ts
.
Fo
r
th
e
liter
atu
r
e
co
m
p
ar
is
o
n
,
ca
n
d
id
ate
s
tu
d
ies
wer
e
s
cr
ee
n
ed
in
two
s
tag
es
(
titl
e/ab
s
tr
ac
t
s
cr
ee
n
in
g
f
o
llo
wed
b
y
f
u
ll
-
tex
t
r
e
v
iew)
.
Stu
d
ies
wer
e
in
clu
d
ed
w
h
e
n
th
ey
:
i
)
ad
d
r
ess
ed
wee
d
d
e
tectio
n
u
s
in
g
d
ee
p
lear
n
in
g
o
r
tr
a
d
itio
n
al
co
m
p
u
ter
v
is
io
n
,
ii)
r
ep
o
r
ted
q
u
a
n
titativ
e
p
er
f
o
r
m
a
n
ce
m
etr
ic
s
(
e.
g
.
,
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
,
F1
-
s
co
r
e
,
an
d
p
r
o
ce
s
s
in
g
tim
e
o
r
FP
S),
an
d
iii)
u
s
ed
ex
p
er
im
e
n
tal
s
ett
in
g
s
co
m
p
ar
ab
le
t
o
f
ield
co
n
d
itio
n
s
.
Data
p
r
o
ce
s
s
in
g
an
d
g
r
o
u
p
in
g
also
in
clu
d
ed
th
e
n
o
r
m
aliza
tio
n
o
f
u
n
its
o
f
m
ea
s
u
r
em
en
t
to
f
ac
ilit
ate
in
ter
-
s
tu
d
y
co
m
p
a
r
is
o
n
s
[
4
7
]
.
E
x
tr
ac
te
d
m
etr
i
cs
wer
e
n
o
r
m
alize
d
wh
er
e
n
ee
d
ed
to
en
a
b
le
tr
an
s
p
ar
en
t c
o
m
p
ar
is
o
n
.
B
r
ief
l
y
,
th
e
p
r
o
ce
s
s
o
f
co
llectin
g
r
esear
ch
d
ata
is
p
r
esen
ted
in
Fig
u
r
e
2
.
Fig
u
r
e
2
.
Data
co
llectio
n
p
r
o
c
ess
2
.
3
.
Da
t
a
a
na
ly
s
is
Qu
an
titativ
e
an
aly
s
is
ass
ess
e
d
th
e
p
er
f
o
r
m
an
ce
o
f
d
ee
p
l
ea
r
n
in
g
-
b
ased
wee
d
d
etec
tio
n
m
eth
o
d
s
(
C
NN,
YOL
Ov
5
,
YOL
O
-
NA
S,
an
d
Ma
s
k
R
-
C
NN)
b
ased
o
n
ac
cu
r
ac
y
an
d
p
r
o
ce
s
s
in
g
s
p
ee
d
.
T
h
e
av
e
r
ag
e
v
alu
e
o
f
ea
c
h
m
etr
ic
was c
alc
u
lated
u
s
in
g
(
1
)
.
̅
=
∑
=
1
(
1
)
W
h
er
e
̅
is
th
e
av
er
ag
e,
is
th
e
in
d
iv
id
u
al
v
al
u
e,
an
d
n
is
th
e
s
am
p
le
s
ize.
E
ac
h
m
eth
o
d
was
co
m
p
iled
f
r
o
m
r
elev
a
n
t
s
tu
d
ies
an
d
th
en
av
er
ag
e
d
to
p
r
o
v
id
e
in
s
ig
h
t
in
to
th
e
o
v
er
all
ef
f
ec
t
iv
en
ess
.
Similar
ly
,
p
r
o
ce
s
s
in
g
s
p
ee
d
was
ev
alu
ated
b
ased
o
n
th
e
r
ep
o
r
te
d
av
er
ag
e
v
alu
e
o
f
p
r
o
ce
s
s
in
g
tim
e
p
er
f
r
am
e
(
m
s
/f
r
am
e)
.
T
h
is
an
aly
s
is
'
s
r
esu
lts
h
elp
ed
id
en
tify
th
e
m
e
th
o
d
th
at
o
f
f
e
r
ed
th
e
b
est
co
m
b
in
atio
n
o
f
h
ig
h
ac
cu
r
ac
y
an
d
tim
e
ef
f
icien
cy
[
4
8
]
.
T
h
e
r
esu
lts
ar
e
v
is
u
alize
d
b
ase
d
o
n
t
h
e
an
aly
ze
d
d
ata
in
g
r
ap
h
s
an
d
tab
les
to
clar
if
y
th
e
in
t
er
-
m
eth
o
d
co
m
p
ar
is
o
n
.
A
s
ca
tter
p
l
o
t
g
r
ap
h
was
u
s
ed
to
s
h
o
w
th
e
r
elatio
n
s
h
ip
b
etwe
en
ac
c
u
r
ac
y
(
%)
an
d
s
p
ee
d
(
m
s
/f
r
am
e)
,
with
th
e
X
-
a
x
is
r
ep
r
esen
tin
g
s
p
ee
d
an
d
th
e
Y
-
ax
is
r
ep
r
esen
tin
g
ac
c
u
r
ac
y
.
T
h
e
Pear
s
o
n
co
r
r
elatio
n
c
o
ef
f
icien
t f
o
r
m
u
la
was u
s
ed
to
ass
ess
th
e
r
elatio
n
s
h
ip
b
etwe
en
th
e
two
v
ar
iab
l
es
as in
(
2
)
.
=
∑
(
−
̅
)
(
−
̅
)
√
∑
(
−
̅
)
2
(
−
̅
)
2
(
2
)
W
h
er
e
an
d
ar
e
th
e
in
d
iv
i
d
u
a
l
v
alu
es
f
o
r
s
p
ee
d
a
n
d
ac
c
u
r
a
cy
,
an
d
̅
,
̅
is
th
e
av
er
ag
e
o
f
ea
ch
v
ar
iab
le.
T
h
is
co
r
r
elatio
n
p
r
o
v
i
d
es
in
s
ig
h
t
in
to
wh
ich
m
eth
o
d
b
alan
c
es
s
p
ee
d
an
d
ac
c
u
r
ac
y
.
T
h
e
v
is
u
aliza
tio
n
r
esu
lts
ar
e
also
d
is
p
lay
ed
in
a
co
m
p
a
r
is
o
n
tab
le
f
o
r
a
m
o
r
e
s
tr
aig
h
tf
o
r
war
d
in
ter
p
r
etatio
n
.
T
h
e
s
tatis
tica
l
m
atr
ices
(
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
s
en
s
itiv
ity
,
an
d
F1
-
s
co
r
e)
wer
e
u
s
ed
to
ev
alu
ate
ea
ch
m
eth
o
d
'
s
p
er
f
o
r
m
a
n
ce
.
T
h
e
m
ath
em
atica
l e
q
u
atio
n
s
f
o
r
ea
ch
m
etr
ic
ar
e
as
(
3
)
to
(
6
)
.
=
+
+
+
+
(
3
)
=
+
(
4
)
=
+
(
5
)
1
−
=
2
×
×
+
(
6
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
4
2
8
-
1
4
4
0
1432
Her
e,
th
e
c
o
n
f
u
s
i
o
n
m
at
r
i
x
p
r
o
v
i
d
es
ess
en
tia
l
m
et
r
ics
,
i
n
cl
u
d
i
n
g
tr
u
e
p
o
s
i
ti
v
e
(
T
P
)
,
tr
u
e
n
eg
ati
v
e
(
T
N)
,
f
als
e
p
o
s
iti
v
e
(
FP
)
,
a
n
d
f
alse
n
e
g
a
tiv
e
(
FN
)
,
wh
ic
h
a
r
e
u
s
e
d
t
o
ca
lc
u
la
te
a
cc
u
r
ac
y
,
p
r
e
cisi
o
n
,
s
e
n
s
it
iv
it
y
,
a
n
d
s
p
e
ci
f
ici
ty
.
D
ata
we
r
e
an
al
y
z
ed
u
s
i
n
g
a
co
n
f
u
s
i
o
n
m
a
tr
ix
to
d
e
te
r
m
i
n
e
a
cc
u
r
ac
y
,
s
p
ee
d
o
f
d
ete
cti
o
n
,
a
n
d
ad
ap
ta
b
il
it
y
to
t
r
o
p
ic
al
e
n
v
i
r
o
n
m
e
n
tal
co
n
d
it
io
n
s
s
u
c
h
as
i
n
Me
r
a
u
k
e.
T
h
e
e
v
al
u
ati
o
n
r
esu
l
ts
w
ill
b
e
c
o
m
p
ar
e
d
wit
h
t
r
a
d
i
ti
o
n
al
m
et
h
o
d
s
t
o
ass
ess
t
h
e
s
u
p
e
r
i
o
r
it
y
o
f
d
e
ep
l
ea
r
n
i
n
g
-
b
as
e
d
m
o
d
els
in
e
f
f
ic
ie
n
c
y
a
n
d
a
cc
u
r
ac
y
.
T
h
e
r
esu
lts
we
r
e
i
n
te
r
p
r
e
te
d
b
y
i
d
e
n
ti
f
y
i
n
g
ea
ch
m
et
h
o
d
'
s
p
e
r
f
o
r
m
a
n
ce
p
a
tte
r
n
s
b
ase
d
o
n
q
u
a
n
ti
tat
iv
e
d
at
a.
Me
t
h
o
d
s
th
at
d
e
m
o
n
s
t
r
at
ed
h
i
g
h
ac
cu
r
a
c
y
(
>9
0
%
)
a
n
d
e
f
f
ic
ie
n
t
p
r
o
c
ess
i
n
g
s
p
ee
d
(
<
5
0
m
s
/
f
r
a
m
e
)
wer
e
co
n
s
id
er
ed
s
u
p
er
io
r
in
th
e
c
o
n
tex
t
o
f
r
ea
l
-
tim
e
wee
d
d
etec
tio
n
ap
p
licatio
n
s
.
I
n
ad
d
itio
n
,
th
e
s
tr
en
g
th
s
an
d
wea
k
n
ess
es
o
f
ea
ch
m
eth
o
d
wer
e
cr
itically
an
aly
ze
d
,
in
cl
u
d
in
g
Ma
s
k
R
-
C
NN
'
s
ab
ilit
y
to
d
etec
t
wee
d
s
with
co
m
p
lex
s
h
a
p
es
b
u
t
at
a
lo
we
r
s
p
ee
d
th
a
n
YOL
Ov
5
,
w
h
ich
is
f
aster
b
u
t
with
s
lig
h
tly
lo
wer
ac
cu
r
ac
y
.
T
h
e
an
aly
s
is
also
ev
alu
ated
th
e
e
n
v
ir
o
n
m
en
tal
s
u
s
tain
ab
ilit
y
i
m
p
licatio
n
s
,
g
iv
e
n
th
at
f
aster
an
d
m
o
r
e
ac
cu
r
ate
m
eth
o
d
s
ca
n
r
ed
u
ce
r
elian
ce
o
n
ch
em
ical
h
er
b
icid
es
[
4
9
]
,
[
5
0
]
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
3
.
1
.
I
nte
rpre
t
a
t
io
n
o
f
r
esu
lt
T
h
e
r
esu
lts
o
f
th
is
s
tu
d
y
c
o
n
f
ir
m
th
e
s
u
p
e
r
io
r
ity
o
f
d
ee
p
lear
n
in
g
m
eth
o
d
s
o
v
er
tr
ad
itio
n
al
tech
n
iq
u
es
in
wee
d
d
etec
tio
n
,
esp
ec
ially
in
th
e
co
n
tex
t
o
f
s
u
s
tain
ab
le
ag
r
icu
ltu
r
e.
Fo
cu
s
in
g
o
n
th
r
ee
m
ai
n
d
ee
p
lear
n
in
g
m
o
d
els,
YOL
O
-
NAS
an
d
Ma
s
k
R
-
C
NN
-
th
is
,
th
is
s
tu
d
y
th
o
r
o
u
g
h
ly
ev
alu
at
es
k
ey
p
er
f
o
r
m
a
n
ce
m
etr
ics:
d
etec
tio
n
ac
cu
r
ac
y
,
p
r
o
ce
s
s
in
g
s
p
ee
d
,
an
d
ad
a
p
ta
b
ilit
y
to
tr
o
p
ical
en
v
ir
o
n
m
en
t
s
s
u
ch
as
Me
r
au
k
e,
I
n
d
o
n
esia.
T
h
ese
f
in
d
in
g
s
ar
e
h
ig
h
ly
r
elev
a
n
t,
as
ef
f
ec
tiv
e
wee
d
d
etec
tio
n
is
o
n
e
o
f
th
e
k
ey
f
ac
to
r
s
in
s
u
p
p
o
r
tin
g
s
u
s
tain
ab
le
cr
o
p
p
r
o
d
u
ctiv
ity
a
n
d
en
v
ir
o
n
m
en
tal
co
n
s
er
v
atio
n
.
T
h
e
YOL
Ov
5
m
et
h
o
d
s
h
o
ws
ex
ce
l
le
n
c
e
in
r
e
al
-
ti
m
e
a
p
p
li
ca
tio
n
s
wi
th
a
d
e
tec
ti
o
n
ac
cu
r
a
c
y
o
f
9
3
.
2
%
an
d
a
v
er
ag
e
i
n
f
e
r
e
n
c
e
ti
m
e
o
f
o
n
l
y
1
2
m
s
p
e
r
i
m
a
g
e
.
T
h
is
m
a
k
es
it
i
d
e
al
f
o
r
la
r
g
e
-
s
ca
le
p
r
ec
is
i
o
n
a
g
r
i
c
u
lt
u
r
e
m
a
n
a
g
e
m
e
n
t
,
w
h
ic
h
r
e
q
u
ir
es
h
ig
h
p
r
o
ce
s
s
i
n
g
s
p
ee
d
wit
h
o
u
t
s
ac
r
i
f
ic
in
g
a
cc
u
r
ac
y
.
O
n
th
e
o
th
er
h
a
n
d
,
YOL
O
-
NAS
y
ie
ld
ed
t
h
e
h
ig
h
e
s
t
a
cc
u
r
ac
y
o
f
9
5
.
6
%
,
m
a
k
i
n
g
it
e
x
ce
ll
e
n
t
ch
o
ice
f
o
r
w
ee
d
d
e
tect
io
n
i
n
c
o
m
p
le
x
s
ce
n
a
r
i
o
s
,
al
b
ei
t
wit
h
t
h
e
c
o
m
p
r
o
m
is
e
o
f
a
s
l
o
w
e
r
p
r
o
c
ess
i
n
g
s
p
ee
d
o
f
2
5
m
s
p
er
im
ag
e.
Ma
s
k
R
-
C
NN
,
wi
th
a
s
eg
m
e
n
t
ati
o
n
ac
c
u
r
a
cy
o
f
9
4
.
1
%,
s
t
a
n
d
s
o
u
t
i
n
p
r
e
cisi
o
n
b
u
t
f
a
ce
s
s
p
ee
d
lim
ita
ti
o
n
s
wi
th
a
n
in
f
e
r
e
n
ce
ti
m
e
o
f
3
1
m
s
p
e
r
i
m
a
g
e
,
m
a
k
i
n
g
it
m
o
r
e
s
u
i
ta
b
l
e
f
o
r
tas
k
s
t
h
at
r
e
q
u
i
r
e
a
d
v
a
n
c
ed
o
b
je
ct
s
e
g
m
en
tat
i
o
n
[
5
1
]
.
I
n
co
n
tr
ast,
tr
ad
itio
n
al
m
eth
o
d
s
o
n
ly
ac
h
iev
e
a
n
av
er
a
g
e
ac
cu
r
ac
y
o
f
8
5
%
with
a
m
u
ch
s
lo
wer
p
r
o
ce
s
s
in
g
tim
e
o
f
m
o
r
e
th
an
5
0
m
s
p
er
im
ag
e
.
T
h
ese
lim
itatio
n
s
em
p
h
asize
th
e
n
ee
d
f
o
r
m
o
r
e
ef
f
icien
c
y
an
d
p
r
ec
is
io
n
f
o
r
lar
g
e
s
ca
les,
es
p
ec
ially
in
co
m
p
lex
tr
o
p
ical
en
v
ir
o
n
m
en
ts
s
u
ch
as
Me
r
au
k
e.
T
h
ese
f
in
d
in
g
s
s
u
p
p
o
r
t th
e
ar
g
u
m
en
t th
at
d
ee
p
lear
n
in
g
m
eth
o
d
s
n
o
t o
n
ly
o
v
er
co
m
e
th
e
ch
allen
g
es o
f
s
p
ee
d
an
d
ac
cu
r
ac
y
b
u
t
also
p
r
o
v
id
e
b
etter
a
d
ap
tab
ilit
y
to
f
ield
co
n
d
itio
n
s
,
ac
co
r
d
in
g
to
p
r
ev
io
u
s
r
esear
ch
f
i
n
d
in
g
s
[
5
1
]
.
T
h
u
s
,
u
s
in
g
d
ee
p
lear
n
in
g
-
b
ased
tec
h
n
o
lo
g
ies
ca
n
b
e
a
p
r
ac
tical
an
d
s
tr
ateg
ic
s
o
lu
tio
n
to
im
p
r
o
v
e
wee
d
m
an
ag
em
en
t
ef
f
icien
cy
i
n
th
e
ag
r
icu
ltu
r
al
s
ec
to
r
.
T
h
e
d
is
tr
ib
u
tio
n
o
f
wee
d
d
etec
tio
n
m
eth
o
d
s
b
as
ed
o
n
a
r
ec
ap
o
f
s
ec
o
n
d
ar
y
d
ata
th
at
d
is
cu
s
s
es we
ed
d
etec
tio
n
m
eth
o
d
s
an
d
tr
ad
itio
n
al
m
eth
o
d
s
is
p
r
esen
ted
in
T
ab
le
1
.
T
ab
le
1
.
Dis
tr
ib
u
tio
n
ta
b
le
o
f
wee
d
d
etec
tio
n
m
eth
o
d
s
No
.
M
o
d
e
l
F
r
e
q
u
e
n
c
y
P
e
r
c
e
n
t
a
g
e
1
Tr
a
d
i
t
i
o
n
a
l
17
1
7
.
0
2
Y
O
LO
-
NAS
24
2
4
.
0
3
M
a
s
k
R
-
C
N
N
26
2
6
.
0
4
Y
O
LO
v
5
33
3
3
.
0
B
ased
o
n
T
ab
le
1
,
th
e
YOL
Ov
5
m
eth
o
d
is
th
e
m
o
s
t
f
r
e
q
u
en
tl
y
u
s
ed
m
o
d
el
in
th
e
liter
atu
r
e,
f
o
llo
wed
b
y
Ma
s
k
R
-
C
NN
.
T
h
is
r
ef
lects
th
e
r
esear
ch
co
m
m
u
n
ity
'
s
p
r
ef
er
e
n
ce
f
o
r
m
o
d
els
k
n
o
wn
f
o
r
t
h
eir
s
p
ee
d
an
d
f
lex
ib
ilit
y
.
Fre
q
u
e
n
cy
v
i
s
u
aliza
tio
n
an
d
p
r
esen
tatio
n
o
f
th
e
d
etec
tio
n
m
o
d
el
s
h
if
t
f
r
o
m
tr
ad
itio
n
al
to
YOL
O
-
NAS,
R
-
C
NN
Ma
s
k
,
a
n
d
YOL
Ov
5
ar
e
p
r
esen
ted
i
n
Fig
u
r
e
3
.
T
h
e
b
a
r
g
r
ap
h
in
Fig
u
r
e
3
s
h
o
ws
ea
ch
m
o
d
el'
s
ab
s
o
lu
te
f
r
eq
u
e
n
cy
o
f
u
s
e,
w
h
ile
th
e
lin
e
g
r
a
p
h
d
is
p
lay
s
its
r
elativ
e
p
er
ce
n
tag
e.
YOL
Ov
5
is
th
e
m
o
s
t
d
o
m
in
an
t
m
o
d
el
at
3
3
%,
r
ef
lectin
g
its
ef
f
icien
cy
an
d
r
elev
an
ce
in
m
o
d
e
r
n
ap
p
licati
o
n
s
.
Ma
s
k
R
-
C
NN
an
d
YOL
O
-
NAS
s
h
o
w
s
ig
n
if
ican
t
co
m
p
etitio
n
,
r
ef
lectin
g
th
eir
u
tili
ty
in
s
ce
n
a
r
io
s
th
at
r
eq
u
ir
e
h
ig
h
p
r
ec
is
io
n
o
r
t
h
e
ab
ilit
y
to
h
an
d
le
co
m
p
le
x
w
ee
d
ch
ar
ac
ter
is
tics
.
Me
an
wh
ile,
with
o
n
ly
1
7
%
o
f
r
ep
r
esen
tatio
n
s
,
tr
a
d
itio
n
al
m
eth
o
d
s
co
n
f
ir
m
th
e
d
ec
r
ea
s
ed
r
elian
ce
o
n
co
n
v
en
tio
n
al
tech
n
iq
u
es o
v
er
d
ee
p
lear
n
in
g
-
b
ased
s
o
l
u
tio
n
s
.
T
h
is
is
a
co
m
p
ar
is
o
n
g
r
a
p
h
o
f
ac
cu
r
ac
y
b
ased
o
n
s
ec
o
n
d
ar
y
d
ata
f
o
u
n
d
f
o
r
ea
c
h
wee
d
d
etec
tio
n
m
o
d
el.
Fig
u
r
e
4
s
h
o
ws
th
at
YOL
O
-
NAS
h
as
th
e
h
ig
h
est
r
an
g
e
o
f
ac
cu
r
ac
y
with
th
e
h
ig
h
est
m
ax
im
u
m
v
alu
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dro
n
e
-
a
s
s
is
ted
d
ee
p
lea
r
n
in
g
w
ee
d
d
etec
tio
n
fo
r
s
u
s
ta
in
a
b
le
a
g
r
icu
ltu
r
e
a
n
d
…
(
A
g
u
s
ta
n
La
tif
)
1433
o
f
9
6
%,
in
d
icatin
g
its
s
u
p
er
io
r
ity
in
d
etec
tin
g
h
ig
h
-
co
m
p
lex
i
ty
wee
d
s
.
Me
an
wh
ile,
t
h
e
tr
ad
itio
n
al
m
eth
o
d
h
as
th
e
lo
west m
ed
ian
ac
cu
r
ac
y
,
r
ein
f
o
r
cin
g
its
wea
k
n
ess
in
lar
g
e
-
s
ca
le
ap
p
licatio
n
s
.
Fig
u
r
e
3
.
Vis
u
aliza
tio
n
o
f
f
r
e
q
u
en
cy
a
n
d
p
e
r
ce
n
tag
e
Fig
u
r
e
4
.
C
o
m
p
a
r
is
o
n
o
f
ac
cu
r
ac
y
b
ased
o
n
m
eth
o
d
s
T
ab
le
2
an
d
Fig
u
r
e
5
s
h
o
w
t
h
e
r
elatio
n
s
h
i
p
b
etwe
en
ac
cu
r
ac
y
an
d
p
r
o
ce
s
s
in
g
tim
e
f
o
r
ea
ch
wee
d
d
etec
tio
n
m
eth
o
d
b
ased
o
n
s
e
co
n
d
ar
y
d
ata.
YOL
Ov
5
h
as
9
4
%
ac
cu
r
ac
y
a
n
d
1
2
m
s
p
r
o
c
ess
in
g
tim
e,
wh
ich
m
ak
es
it
id
ea
l
f
o
r
r
ea
l
-
tim
e
ap
p
licatio
n
s
.
YOL
O
-
NAS
h
as
th
e
h
ig
h
est
ac
c
u
r
ac
y
o
f
9
6
%
an
d
a
p
r
o
ce
s
s
in
g
tim
e
o
f
2
5
m
s
,
s
u
itab
le
f
o
r
co
m
p
lex
s
ce
n
ar
io
s
.
Ma
s
k
R
-
C
NN
h
a
s
9
4
% a
cc
u
r
ac
y
,
3
1
m
s
p
r
o
ce
s
s
in
g
tim
e,
an
d
e
x
ce
ls
in
s
eg
m
en
tatio
n
p
r
ec
is
io
n
.
T
h
e
tr
ad
itio
n
al
m
eth
o
d
h
as
8
5
%
ac
cu
r
ac
y
a
n
d
a
p
r
o
ce
s
s
in
g
tim
e
o
f
>5
0
m
s
,
lim
itin
g
its
s
ca
lab
ilit
y
.
T
h
e
tr
e
n
d
lin
e
o
n
th
e
g
r
ap
h
s
h
o
ws
a
s
lig
h
t
in
cr
ea
s
e,
w
h
ich
in
d
icate
s
a
m
in
o
r
t
r
ad
e
-
o
f
f
,
m
ea
n
in
g
m
o
d
els
with
h
ig
h
er
a
cc
u
r
ac
y
ten
d
to
r
eq
u
ir
e
lo
n
g
er
p
r
o
ce
s
s
in
g
tim
e.
T
ab
le
3
s
h
o
w
s
th
e
d
etails
o
f
th
e
s
tatis
t
ical
an
aly
s
is
.
T
ab
le
2
.
Acc
u
r
ac
y
an
d
p
r
o
ce
s
s
in
g
tim
e
r
elatio
n
s
h
ip
d
ata
N
o
.
M
o
d
e
l
A
c
c
u
r
a
c
y
(
%)
P
r
o
c
e
ss
i
n
g
t
i
me
(
ms)
1
Y
O
LO
v
5
9
4
.
0
12
2
Y
O
LO
-
NAS
9
6
.
0
25
3
M
a
s
k
R
-
C
N
N
9
3
.
1
31
4
Tr
a
d
i
t
i
o
n
a
l
8
5
.
0
50
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
4
2
8
-
1
4
4
0
1434
Fig
u
r
e
5
.
R
elatio
n
s
h
ip
b
etwe
e
n
ac
cu
r
ac
y
an
d
p
r
o
ce
s
s
in
g
tim
e
T
ab
le
3
.
Deta
iled
s
tatis
tical
an
aly
s
is
N
o
.
M
a
t
r
i
c
V
a
l
u
e
1
M
e
a
n
a
c
c
u
r
a
c
y
9
0
.
3
5
2
M
e
a
n
p
r
o
c
e
ssi
n
g
t
i
me
(
ms)
2
9
.
5
0
3
P
e
r
so
n
c
o
r
r
e
l
a
t
i
o
n
-
0
.
8
2
4
R
-
sq
u
a
r
e
(
l
i
n
e
a
r
r
e
g
r
e
ss
i
o
n
)
0
.
6
8
5
P
-
v
a
l
u
e
(
l
i
n
e
a
r
r
e
g
r
e
ss
i
o
n
)
0
.
1
8
6
S
t
a
n
d
a
r
d
e
r
r
o
r
(
l
i
n
e
a
r
r
e
g
r
e
ssi
o
n
)
0
.
8
0
T
ab
le
3
s
h
o
ws
th
at
f
o
r
all
m
o
d
els,
th
e
av
er
a
g
e
ac
cu
r
ac
y
is
9
0
.
3
5
%,
an
d
th
e
av
e
r
ag
e
p
r
o
c
ess
in
g
tim
e
is
2
9
.
5
0
m
s
p
er
im
a
g
e.
T
h
e
P
ea
r
s
o
n
test
(
p
=
0
.
1
8
)
s
h
o
ws
n
o
s
tatis
tically
s
ig
n
if
ican
t
lin
ea
r
r
elatio
n
s
h
ip
at
a
s
ig
n
if
ican
ce
lev
el
o
f
α
=0
.
0
5
.
L
in
ea
r
r
eg
r
ess
io
n
y
ield
s
R
²=
0
.
6
8
(
6
7
.
8
2
%
v
a
r
ian
ce
ex
p
lain
ed
)
with
a
s
tan
d
ar
d
er
r
o
r
o
f
0
.
8
0
,
i
n
d
icatin
g
m
o
d
e
r
ate
f
it.
T
h
e
r
elatio
n
s
h
ip
b
etwe
en
ac
c
u
r
ac
y
a
n
d
p
r
o
ce
s
s
in
g
tim
e
d
i
s
p
lay
s
a
m
ar
g
i
n
al
tr
ad
e
-
o
f
f
:
a
s
ac
cu
r
ac
y
in
cr
ea
s
es,
p
r
o
ce
s
s
in
g
tim
e
ten
d
s
to
r
is
e
s
lig
h
tly
,
th
o
u
g
h
th
is
ass
o
ciatio
n
is
n
o
t
s
tatis
tically
s
ig
n
if
ican
t.
Fu
r
th
er
r
esear
ch
is
n
ee
d
e
d
to
o
p
tim
ize
b
o
th
m
etr
ics
co
n
cu
r
r
e
n
tly
.
Y
OL
Ov
5
r
em
ain
s
p
r
ef
e
r
r
ed
f
o
r
r
ea
l
-
tim
e
u
s
e
d
u
e
to
its
s
p
ee
d
,
wh
ile
YOL
O
-
NAS
a
n
d
Ma
s
k
R
-
C
NN
ex
ce
l in
s
ce
n
ar
io
s
p
r
io
r
itizin
g
h
ig
h
er
ac
cu
r
ac
y
.
Per
f
o
r
m
an
ce
r
esu
lts
f
r
o
m
a
s
ec
o
n
d
ar
y
an
al
y
s
is
o
f
1
0
0
jo
u
r
n
als
o
n
wee
d
d
etec
tio
n
m
et
h
o
d
s
u
s
in
g
d
ee
p
lear
n
in
g
an
d
tr
ad
itio
n
al
t
ec
h
n
iq
u
es
ar
e
p
r
esen
ted
in
T
a
b
le
4
.
T
h
is
tab
le
s
h
o
ws
th
at
d
ee
p
lear
n
in
g
-
b
ased
m
eth
o
d
s
ac
h
iev
e
s
ig
n
if
ica
n
t
p
er
f
o
r
m
a
n
ce
in
wee
d
d
etec
tio
n
.
YOL
O
-
NAS
d
is
p
lay
ed
t
h
e
h
ig
h
est
ac
cu
r
ac
y
a
n
d
r
ec
all,
m
ak
in
g
it th
e
s
u
p
er
io
r
m
eth
o
d
f
o
r
th
is
co
m
p
ar
is
o
n
.
T
h
is
is
m
ain
ly
d
u
e
to
th
e
m
o
d
el
ar
ch
itectu
r
e,
wh
ich
ca
n
ca
p
tu
r
e
co
m
p
lex
f
ea
tu
r
es m
o
r
e
ef
f
ec
tiv
ely
.
Me
an
w
h
ile,
tr
ad
itio
n
al
tech
n
iq
u
es
p
er
f
o
r
m
ed
wo
r
s
e
th
an
d
ee
p
lear
n
in
g
m
eth
o
d
s
.
T
h
is
r
ef
lect
s
th
e
lim
itatio
n
s
in
h
an
d
lin
g
im
ag
e
co
m
p
lex
ity
o
r
v
a
r
ied
tr
o
p
ical
en
v
ir
o
n
m
en
ts
s
u
ch
as
Me
r
a
u
k
e.
Nex
t,
th
e
r
esu
lts
o
f
th
e
p
e
r
f
o
r
m
an
ce
m
a
tr
ix
co
m
p
ar
is
o
n
(
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e)
f
o
r
v
ar
i
o
u
s
wee
d
d
et
ec
tio
n
m
eth
o
d
s
ar
e
in
ter
p
r
eted
.
Fig
u
r
e
6
s
u
m
m
a
r
izes
th
e
tr
e
n
d
in
p
er
f
o
r
m
an
ce
im
p
r
o
v
e
m
en
t
f
r
o
m
tr
ad
itio
n
al
t
o
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
es.
T
h
is
f
ig
u
r
e
is
a
g
r
ap
h
o
f
th
e
tr
en
d
in
p
er
f
o
r
m
a
n
ce
im
p
r
o
v
e
m
en
t
f
r
o
m
tr
ad
itio
n
al
tech
n
iq
u
es
to
d
ee
p
lear
n
i
n
g
-
b
ased
m
eth
o
d
s
.
T
h
is
in
d
icate
s
th
at
m
o
d
er
n
m
eth
o
d
s
a
r
e
r
ec
o
m
m
en
d
ed
f
o
r
ac
cu
r
ate
a
n
d
ef
f
icien
t
wee
d
d
etec
tio
n
.
YOL
O
-
NAS
o
f
f
er
s
th
e
b
est
p
er
f
o
r
m
an
ce
,
esp
ec
ially
in
ter
m
s
o
f
ac
cu
r
ac
y
an
d
r
ec
all,
wh
ich
ar
e
ess
en
tial
f
o
r
en
v
ir
o
n
m
en
tal
s
u
s
tain
ab
ilit
y
.
Ho
wev
er
,
all
d
ee
p
lear
n
in
g
-
b
ased
m
eth
o
d
s
p
r
o
v
i
d
e
s
ig
n
if
ican
t a
d
v
an
ta
g
es o
v
er
tr
a
d
itio
n
al
tech
n
iq
u
es,
b
o
th
i
n
ef
f
icien
cy
an
d
e
f
f
ec
tiv
en
ess
.
T
ab
le
4
.
Per
f
o
r
m
an
ce
r
esu
lts
o
f
wee
d
d
etec
tio
n
m
eth
o
d
M
o
d
e
l
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
s
c
o
r
e
(
%)
P
r
o
c
e
ss
i
n
g
s
p
e
e
d
(
f
p
s)
Y
O
LO
v
5
94
92
93
93
45
Y
O
LO
-
NAS
96
94
95
94
50
M
a
s
k
R
-
C
N
N
93
91
92
91
30
Tr
a
d
i
t
i
o
n
a
l
85
80
83
81
10
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dro
n
e
-
a
s
s
is
ted
d
ee
p
lea
r
n
in
g
w
ee
d
d
etec
tio
n
fo
r
s
u
s
ta
in
a
b
le
a
g
r
icu
ltu
r
e
a
n
d
…
(
A
g
u
s
ta
n
La
tif
)
1435
Fig
u
r
e
6
.
Per
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
o
f
wee
d
d
etec
tio
n
m
eth
o
d
3
.
2
.
I
m
pa
c
t
o
n
t
heo
r
y
a
nd
pra
ct
ice
T
h
is
s
tu
d
y
'
s
co
n
tr
ib
u
tio
n
to
th
eo
r
y
d
ev
el
o
p
m
en
t
is
ess
en
tial
to
d
e
v
elo
p
in
g
d
ee
p
lea
r
n
in
g
-
b
ased
o
b
ject
d
etec
tio
n
th
eo
r
y
,
esp
ec
ially
in
th
e
co
n
tex
t
o
f
p
r
ec
is
io
n
ag
r
icu
ltu
r
e
in
tr
o
p
ical
r
eg
io
n
s
s
u
ch
as
Me
r
au
k
e.
T
h
is
s
tu
d
y
r
ev
ea
ls
th
e
s
u
p
er
i
o
r
ity
o
f
d
ee
p
lear
n
in
g
m
o
d
els
in
o
v
er
co
m
in
g
wee
d
d
etec
t
io
n
ch
allen
g
es
b
y
co
m
p
ar
in
g
th
e
YOL
Ov
5
m
o
d
el,
YOL
O
-
NAS,
Ma
s
k
R
-
C
N
N,
an
d
tr
a
d
itio
n
al
m
et
h
o
d
s
.
T
h
e
r
esu
lts
s
h
o
w
th
at
YOL
Ov
5
ca
n
ac
h
iev
e
9
4
%
a
cc
u
r
ac
y
with
a
p
r
o
ce
s
s
in
g
ti
m
e
o
f
o
n
ly
1
2
m
s
,
m
ak
in
g
it
id
ea
l
f
o
r
r
ea
l
-
tim
e
ap
p
licatio
n
s
.
T
h
ese
a
d
v
an
tag
es
ar
e
p
ar
ticu
lar
l
y
r
elev
a
n
t
in
Me
r
au
k
e
,
wh
er
e
m
an
u
al
wee
d
co
n
tr
o
l
is
a
s
ig
n
if
ican
t
ch
allen
g
e
th
at
in
c
r
ea
s
es
p
r
o
d
u
ctio
n
co
s
ts
.
W
ith
its
v
ast
ar
ea
s
o
f
f
er
tile
la
n
d
,
Me
r
au
k
e
f
ac
es
t
h
e
p
r
o
b
lem
o
f
co
m
p
etitio
n
b
etwe
en
wee
d
s
an
d
m
ain
cr
o
p
s
f
o
r
n
u
tr
ien
ts
,
wate
r
an
d
lig
h
t.
T
h
e
f
in
d
in
g
s
o
f
t
h
is
r
esear
ch
ex
ten
d
th
e
th
eo
r
y
o
f
wee
d
d
etec
tio
n
m
o
d
el
o
p
tim
izatio
n
b
y
s
h
o
win
g
h
o
w
d
ee
p
lear
n
in
g
al
g
o
r
ith
m
s
ca
n
b
e
ad
ap
ted
f
o
r
co
m
p
le
x
tr
o
p
ical
en
v
ir
o
n
m
e
n
ts
[
5
2
]
,
[
5
3
]
.
T
h
e
co
r
r
elatio
n
b
et
wee
n
ac
cu
r
ac
y
an
d
p
r
o
ce
s
s
in
g
tim
e
s
h
o
wn
in
th
i
s
s
tu
d
y
(
-
0
.
8
2
)
s
u
p
p
o
r
ts
th
e
d
ev
elo
p
m
e
n
t
o
f
e
f
f
icien
cy
-
a
n
d
ac
cu
r
ac
y
-
b
ased
d
etec
tio
n
th
eo
r
ies
u
r
g
e
n
tly
n
ee
d
ed
f
o
r
lar
g
e
-
s
ca
le
ag
r
icu
ltu
r
e
in
Me
r
au
k
e.
Mo
d
els
s
u
ch
as
YOL
Ov
5
an
d
YOL
O
-
NAS
ca
n
b
e
th
e
b
asis
f
o
r
f
u
r
th
er
d
ev
elo
p
i
n
g
tr
o
p
ical
ag
r
icu
ltu
r
al
tech
n
o
lo
g
y
th
eo
r
y
.
T
h
e
p
r
ac
tical
im
p
licatio
n
s
o
f
t
h
is
r
esear
ch
ar
e
s
ig
n
if
ican
t,
esp
ec
ially
in
s
u
p
p
o
r
tin
g
Me
r
au
k
e'
s
ef
f
o
r
ts
as
a
n
atio
n
al
f
o
o
d
p
r
o
d
u
ctio
n
ce
n
ter
.
W
ith
YOL
Ov
5
'
s
h
ig
h
ac
cu
r
ac
y
(
9
4
%)
an
d
f
ast
p
r
o
ce
s
s
in
g
tim
e
(
1
2
m
s
)
,
th
is
tech
n
o
lo
g
y
o
f
f
er
s
an
in
n
o
v
ativ
e
s
o
lu
tio
n
f
o
r
wee
d
d
e
tectio
n
ac
r
o
s
s
Me
r
au
k
e'
s
v
ast
ag
r
icu
ltu
r
al
la
n
d
.
W
ee
d
s
th
at
ar
e
a
s
ig
n
if
ican
t c
o
n
s
tr
ain
t c
an
b
e
id
en
tifie
d
an
d
co
n
tr
o
lled
m
o
r
e
ef
f
icien
tly
,
r
e
d
u
cin
g
th
e
n
ee
d
f
o
r
m
an
u
al
lab
o
r
,
w
h
ich
h
as
b
ee
n
a
m
aj
o
r
c
o
n
tr
ib
u
t
o
r
to
h
ig
h
p
r
o
d
u
ctio
n
co
s
ts
.
Fu
r
th
er
m
o
r
e,
m
o
d
els
s
u
ch
as
YOL
O
-
NAS
an
d
Ma
s
k
R
-
C
NN,
with
9
6
%
an
d
9
3
%
ac
c
u
r
a
cy
,
ca
n
b
e
ap
p
lied
to
m
o
r
e
co
m
p
lex
a
g
r
icu
ltu
r
al
s
ce
n
ar
io
s
in
Me
r
au
k
e,
in
clu
d
in
g
lan
d
with
to
p
o
g
r
a
p
h
ical
v
a
r
iatio
n
s
an
d
d
if
f
e
r
en
t
cr
o
p
ty
p
es.
Ap
p
ly
in
g
th
ese
tech
n
o
lo
g
ies
r
ed
u
ce
s
p
r
o
d
u
cti
o
n
co
s
ts
an
d
in
cr
ea
s
es
cr
o
p
p
r
o
d
u
ctiv
ity
b
y
en
s
u
r
in
g
th
at
k
ey
cr
o
p
s
r
ec
eiv
e
o
p
tim
al
ac
ce
s
s
to
n
u
tr
ien
ts
,
w
ater
,
an
d
lig
h
t.
Ov
er
all,
tar
g
eted
d
etec
tio
n
ca
n
im
p
r
o
v
e
ac
ce
s
s
to
cr
o
p
r
eso
u
r
ce
s
an
d
s
u
p
p
o
r
t
s
u
s
tain
ab
le
m
a
n
ag
em
en
t.
T
h
ese
f
in
d
in
g
s
s
u
p
p
o
r
t
th
e
u
r
g
en
t
n
ee
d
f
o
r
in
n
o
v
ativ
e
lan
d
m
an
ag
em
en
t
ap
p
r
o
ac
h
es
in
Me
r
au
k
e,
as
o
u
tlin
ed
in
th
e
r
ese
ar
ch
b
ac
k
g
r
o
u
n
d
.
Dee
p
lear
n
i
n
g
m
o
d
els
s
u
ch
as
YOL
Ov
5
en
ab
le
ef
f
ec
tiv
e,
e
f
f
icien
t,
an
d
s
u
s
tain
ab
le
wee
d
d
etec
tio
n
,
p
r
o
v
id
in
g
a
s
tr
ateg
ic
s
o
lu
tio
n
f
o
r
m
o
d
er
n
tr
o
p
ical
a
g
r
icu
ltu
r
e
i
n
th
is
r
eg
io
n
.
B
ey
o
n
d
wee
d
d
etec
tio
n
,
th
e
m
eth
o
d
o
lo
g
ical
in
s
ig
h
ts
an
d
em
p
ir
ical
f
in
d
in
g
s
in
th
is
s
tu
d
y
ar
e
r
elev
an
t
f
o
r
a
wid
e
r
an
g
e
o
f
AI
ap
p
lic
atio
n
s
in
ag
r
icu
ltu
r
e
th
at
r
ely
o
n
im
ag
e
an
aly
s
is
in
tr
o
p
ical
en
v
ir
o
n
m
en
ts
.
T
h
e
co
m
b
in
atio
n
o
f
UAV
-
b
ased
im
ag
in
g
,
wee
d
d
atasets
,
an
d
co
m
p
a
r
ativ
e
ev
al
u
atio
n
o
f
cu
ttin
g
-
ed
g
e
d
ee
p
lear
n
in
g
m
o
d
els
p
r
o
v
id
es
a
tr
an
s
f
er
ab
le
an
d
a
d
ap
tab
le
f
r
am
ewo
r
k
f
o
r
r
elate
d
task
s
,
s
u
ch
as
p
est
an
d
d
is
ea
s
e
d
etec
tio
n
,
p
lan
t
h
ea
lth
m
o
n
ito
r
in
g
,
an
d
ea
r
ly
d
ia
g
n
o
s
is
o
f
s
tr
ess
in
m
ajo
r
f
o
o
d
cr
o
p
s
[
5
4
]
,
[
5
5
]
,
wh
ile
s
im
ilar
tr
ain
in
g
ar
ch
itectu
r
es a
n
d
s
tr
at
eg
ies (
in
clu
d
in
g
h
an
d
lin
g
class
im
b
alan
ce
s
an
d
r
esil
ien
ce
to
lig
h
tin
g
v
ar
iatio
n
s
)
ca
n
s
u
p
p
o
r
t
s
y
s
tem
s
in
te
g
r
at
ed
r
esu
lts
th
at
s
im
u
ltan
eo
u
s
l
y
an
aly
ze
wee
d
s
,
p
ests
,
an
d
s
y
m
p
to
m
s
in
p
lan
t
tis
s
u
es
f
r
o
m
th
e
s
am
e
im
ag
e
[
5
6
]
,
[
5
7
]
.
Alth
o
u
g
h
th
is
s
tu
d
y
is
r
o
o
ted
in
th
e
s
p
ec
if
ic
co
n
te
x
t
o
f
Me
r
au
k
e,
th
e
ch
allen
g
es a
d
d
r
ess
ed
s
u
ch
as
cr
o
p
ca
n
o
p
y
h
eter
o
g
en
eity
,
lig
h
tin
g
d
y
n
a
m
ics,
an
d
b
ac
k
g
r
o
u
n
d
co
m
p
lex
ity
s
u
ch
as
s
o
il,
wate
r
,
ar
e
also
co
m
m
o
n
in
m
a
n
y
tr
o
p
ical
ag
r
icu
ltu
r
al
ar
ea
s
,
s
o
th
e
ev
alu
ated
ap
p
r
o
ac
h
h
as
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
4
2
8
-
1
4
4
0
1436
p
o
ten
tial
to
b
e
g
e
n
er
alize
d
to
v
ar
io
u
s
AI
-
b
ased
a
g
r
icu
ltu
r
a
l
m
o
n
ito
r
in
g
task
s
an
d
m
o
r
e
co
m
p
r
eh
e
n
s
iv
e
an
d
d
ata
-
d
r
iv
e
n
cr
o
p
m
an
a
g
em
en
t
s
tr
ateg
ies in
tr
o
p
ical
p
r
o
d
u
ctio
n
s
y
s
tem
s
[
5
8
]
,
[
5
9
]
.
3
.
3
.
Resea
rc
h
l
im
it
a
t
io
ns
T
h
is
s
tu
d
y
u
s
es
th
e
p
r
im
ar
y
d
ataset
o
f
wee
d
s
p
ec
ies
in
Me
r
au
k
e,
o
b
tain
e
d
u
s
in
g
a
d
r
o
n
e
(
UAV)
,
an
d
th
e
s
ec
o
n
d
ar
y
d
ataset,
o
b
tain
e
d
f
r
o
m
s
ev
er
al
s
cien
tific
p
u
b
licatio
n
s
,
is
in
ac
co
r
d
an
ce
wit
h
th
e
wee
d
s
p
ec
ies
o
b
s
er
v
ed
in
th
e
f
ield
.
T
h
is
s
ec
o
n
d
ar
y
d
ataset
is
p
r
im
ar
ily
u
s
ed
f
o
r
q
u
alitativ
e
v
alid
atio
n
an
d
co
n
tex
tu
al
co
m
p
ar
is
o
n
o
f
m
o
d
el
p
er
f
o
r
m
an
ce
,
r
ath
er
th
an
f
o
r
a
cr
o
s
s
-
s
tu
d
y
q
u
an
titativ
e
m
eta
-
an
a
ly
s
is
.
T
h
is
ap
p
r
o
ac
h
p
r
o
v
id
es
d
ata
t
h
at
b
etter
alig
n
s
with
lo
ca
l
ch
ar
ac
ter
is
tics
,
s
u
ch
as
th
e
d
o
m
in
a
n
t
wee
d
ty
p
e,
th
e
p
lan
tin
g
p
atter
n
s
o
f
th
e
lo
ca
l c
o
m
m
u
n
it
y
,
an
d
t
h
e
ty
p
ical
en
v
ir
o
n
m
en
t
al
co
n
d
itio
n
s
in
th
e
Me
r
au
k
e
s
wam
p
an
d
d
r
y
lan
d
ar
ea
s
.
Ho
wev
er
,
th
e
u
s
e
o
f
f
ield
an
d
s
ec
o
n
d
ar
y
d
atasets
s
till
h
as
s
o
m
e
lim
itatio
n
s
.
V
ar
iatio
n
s
in
im
ag
e
q
u
ality
,
d
i
f
f
er
en
ce
s
i
n
lig
h
tin
g
co
n
d
itio
n
s
d
u
r
in
g
d
ata
co
lle
ctio
n
,
an
d
im
b
ala
n
ce
s
in
s
am
p
le
co
u
n
ts
b
etwe
en
wee
d
ty
p
es c
an
af
f
ec
t th
e
ac
cu
r
ac
y
o
f
th
e
d
etec
tio
n
m
o
d
el
a
n
d
p
r
o
ce
s
s
in
g
tim
e
[
6
0
]
.
I
n
ad
d
itio
n
,
alth
o
u
g
h
th
e
d
ataset
u
s
ed
h
as
b
ee
n
ad
ju
s
ted
to
th
e
lo
ca
l
co
n
tex
t,
th
e
s
co
p
e
o
f
o
b
s
er
v
atio
n
is
s
till
lim
ited
to
a
f
ew
s
p
ec
if
ic
l
o
ca
tio
n
s
,
s
o
th
at
it
d
o
es
n
o
t
f
u
lly
r
e
p
r
esen
t
th
e
en
tir
e
h
eter
o
g
en
eity
o
f
Me
r
au
k
e'
s
ag
r
icu
ltu
r
al
ec
o
s
y
s
tem
.
Facto
r
s
s
u
ch
as
th
e
d
iv
er
s
ity
o
f
p
lan
t
v
ar
ieties,
s
ea
s
o
n
al
ch
an
g
es,
a
n
d
th
e
r
ate
o
f
wee
d
in
f
estatio
n
s
in
ea
ch
v
illag
e
c
an
g
iv
e
r
is
e
t
o
ad
d
itio
n
al
v
ar
i
atio
n
s
th
at
h
av
e
n
o
t
b
ee
n
co
m
p
r
e
h
en
s
iv
ely
r
ec
o
r
d
ed
[
6
1
]
,
[
6
2
]
.
T
h
er
ef
o
r
e,
alth
o
u
g
h
th
e
r
esu
lts
o
f
th
is
s
tu
d
y
ca
n
s
u
p
p
o
r
t
th
e
d
ev
elo
p
m
e
n
t
o
f
d
ee
p
lear
n
i
n
g
-
b
ased
wee
d
d
etec
tio
n
tech
n
o
lo
g
y
f
o
r
th
e
tr
o
p
ics,
th
e
f
i
n
d
in
g
s
m
ay
s
till
b
e
b
iased
d
u
e
to
lim
ited
o
b
s
er
v
ati
o
n
co
v
er
ag
e
a
n
d
v
a
r
iatio
n
s
in
en
v
ir
o
n
m
en
tal
co
n
d
itio
n
s
.
T
h
is
r
esear
ch
m
eth
o
d
o
lo
g
y
p
r
io
r
itizes
s
tatis
t
ical
an
aly
s
is
t
o
co
m
p
a
r
e
th
e
ac
cu
r
ac
y
an
d
p
r
o
ce
s
s
in
g
tim
e
o
f
d
ee
p
lear
n
in
g
m
o
d
el
s
s
u
ch
as
YOL
Ov
5
,
YOL
O
-
NAS,
an
d
Ma
s
k
R
-
C
NN.
Al
t
h
o
u
g
h
th
is
an
aly
s
is
p
r
o
v
id
es
a
n
o
v
e
r
v
iew
o
f
th
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
o
n
r
ea
l
d
ata
f
r
o
m
Me
r
au
k
e
,
th
e
s
tu
d
y
d
o
es
n
o
t
in
clu
d
e
f
u
r
th
er
ev
alu
atio
n
o
f
e
x
ter
n
al
ch
an
g
es
s
u
ch
as
e
x
tr
em
e
lig
h
t
in
ten
s
ity
,
s
ea
s
o
n
al
wee
d
g
r
o
wth
d
y
n
am
ics,
o
r
lo
ca
l w
ea
th
er
co
n
d
itio
n
s
th
at
ca
n
s
ig
n
if
ican
tly
af
f
ec
t
m
o
d
el
p
er
f
o
r
m
a
n
ce
[
6
3
]
.
L
in
ea
r
r
eg
r
ess
io
n
an
aly
s
is
u
s
ed
to
ev
alu
ate
th
e
r
elatio
n
s
h
i
p
b
etwe
en
ac
cu
r
ac
y
a
n
d
p
r
o
ce
s
s
in
g
tim
e
also
h
as
lim
itatio
n
s
.
T
h
e
r
elatio
n
s
h
ip
b
etwe
en
th
e
two
v
ar
iab
les
is
lik
ely
n
o
n
-
lin
ea
r
,
esp
ec
ia
lly
s
in
ce
d
ee
p
lear
n
in
g
m
o
d
els
o
f
ten
s
h
o
w
a
d
is
p
r
o
p
o
r
tio
n
ate
in
cr
ea
s
e
in
a
cc
u
r
ac
y
with
th
e
in
c
r
ea
s
ed
co
m
p
u
tatio
n
al
co
m
p
lex
ity
.
T
h
e
R
-
s
q
u
ar
ed
v
alu
e
o
f
6
7
.
8
2
%
in
d
icate
s
th
e
ex
is
ten
c
e
o
f
v
a
r
iab
ilit
y
th
at
h
as
n
o
t
b
ee
n
ex
p
lain
e
d
b
y
th
is
s
im
p
le
lin
ea
r
m
o
d
el.
T
h
u
s
,
f
u
r
th
er
an
al
y
s
is
u
s
in
g
n
o
n
-
lin
ea
r
ap
p
r
o
ac
h
es
o
r
m
o
r
e
c
o
m
p
lex
s
tatis
tical
m
o
d
els
is
n
ee
d
ed
to
illu
s
tr
ate
th
e
tr
u
e
r
elatio
n
s
h
ip
b
etwe
en
ac
cu
r
ac
y
an
d
p
r
o
ce
s
s
in
g
tim
e
in
th
e
co
n
tex
t o
f
wee
d
d
etec
tio
n
in
Me
r
au
k
e.
3
.
4
.
Su
g
g
estio
ns
f
o
r
f
uture
re
s
ea
rc
h
Fu
tu
r
e
r
esear
ch
s
tr
o
n
g
ly
r
e
co
m
m
en
d
s
u
s
in
g
p
r
im
a
r
y
d
ata
co
llected
d
ir
ec
tly
f
r
o
m
tr
o
p
ical
ag
r
icu
ltu
r
al
e
n
v
ir
o
n
m
en
ts
s
u
c
h
as
Me
r
au
k
e,
I
n
d
o
n
esia.
T
h
i
s
aim
s
to
in
cr
ea
s
e
t
h
e
r
ele
v
a
n
ce
o
f
th
e
r
esear
ch
r
esu
lts
to
lo
ca
l
co
n
d
itio
n
s
,
esp
ec
ially
g
iv
en
th
e
wee
d
ch
ar
ac
t
er
is
tics
an
d
en
v
ir
o
n
m
en
tal
v
a
r
iab
ilit
y
in
tr
o
p
ical
r
eg
io
n
s
,
wh
ich
m
ay
d
i
f
f
er
f
r
o
m
th
o
s
e
in
th
e
d
atasets
u
s
ed
in
p
r
ev
io
u
s
s
tu
d
ies.
T
h
ese
p
r
i
m
ar
y
d
atasets
also
en
ab
le
th
e
ev
alu
atio
n
o
f
th
e
p
er
f
o
r
m
an
ce
o
f
d
ee
p
lear
n
in
g
m
o
d
els,
s
u
ch
as
YOL
Ov
5
a
n
d
YOL
O
-
NAS,
o
n
r
ea
l
s
ce
n
ar
io
s
with
ex
te
r
n
al
f
ac
to
r
s
s
u
ch
as
lig
h
t
in
ten
s
it
y
,
wee
d
t
y
p
e
v
ar
iatio
n
,
an
d
lan
d
h
eter
o
g
en
eity
.
Fu
r
th
er
m
o
r
e
,
th
is
an
al
y
s
is
u
s
ed
lin
ea
r
r
e
g
r
ess
io
n
to
ev
alu
ate
th
e
r
elatio
n
s
h
ip
b
etw
ee
n
ac
cu
r
ac
y
an
d
p
r
o
ce
s
s
in
g
tim
e,
wh
ic
h
m
ay
n
o
t
f
u
lly
r
ef
lect
th
e
co
m
p
lex
i
ty
o
f
t
h
e
r
elatio
n
s
h
ip
b
etwe
e
n
th
e
two
v
ar
iab
les.
Fo
r
ex
am
p
le,
th
e
a
n
aly
s
is
r
esu
lts
s
h
o
w
an
R
-
s
q
u
ar
ed
v
alu
e
o
f
6
7
.
8
2
%,
wh
ich
in
d
icate
s
s
i
g
n
if
ican
t
v
ar
ia
b
ilit
y
th
at
a
s
im
p
le
lin
ea
r
m
o
d
el
ca
n
n
o
t
ex
p
lain
.
As
s
u
ch
,
th
is
s
tu
d
y
h
as
lim
itatio
n
s
in
ex
p
lo
r
i
n
g
o
th
er
n
o
n
-
lin
ea
r
o
r
co
m
p
lex
r
elatio
n
s
h
ip
s
th
at
m
a
y
b
e
r
elev
a
n
t f
o
r
d
ev
el
o
p
in
g
m
o
r
e
s
o
p
h
is
ticated
m
o
d
els
[
6
4
]
,
[
6
5
]
.
Fu
r
th
er
m
o
r
e
,
ex
p
an
d
in
g
th
e
d
ataset
'
s
s
co
p
e
to
in
clu
d
e
im
ag
es
with
v
ar
y
in
g
r
eso
lu
tio
n
s
an
d
d
y
n
am
ic
en
v
ir
o
n
m
en
tal
co
n
d
itio
n
s
is
es
s
en
tial.
T
h
is
s
u
p
p
o
r
ts
th
e
d
ev
elo
p
m
en
t
o
f
m
o
r
e
ad
ap
tiv
e
a
n
d
r
o
b
u
s
t
m
o
d
els,
as
p
r
o
p
o
s
ed
b
y
[
66
]
,
i
n
wh
ich
d
ee
p
lear
n
in
g
m
o
d
els
tr
ain
ed
o
n
h
eter
o
g
en
eo
u
s
d
ata
s
h
o
we
d
b
etter
p
er
f
o
r
m
a
n
ce
in
lar
g
e
-
s
ca
le
ag
r
icu
ltu
r
al
a
p
p
licatio
n
s
.
B
y
ex
p
an
d
in
g
t
h
e
d
ataset,
f
u
tu
r
e
r
esear
ch
ca
n
h
e
lp
d
ev
elo
p
m
o
d
els
th
at
ar
e
n
o
t
o
n
ly
lo
ca
ll
y
r
elev
a
n
t b
u
t a
ls
o
g
e
n
er
alize
to
o
t
h
er
tr
o
p
ical
r
eg
io
n
s
.
I
n
ad
d
itio
n
t
o
o
p
tim
izin
g
e
x
is
tin
g
m
o
d
els,
s
u
ch
as
YOL
O
v
5
an
d
Ma
s
k
R
-
C
NN,
f
u
tu
r
e
r
esear
ch
co
u
ld
ex
p
l
o
r
e
n
ew
m
eth
o
d
s
,
in
clu
d
in
g
v
is
io
n
tr
an
s
f
o
r
m
e
r
s
(
ViT
s
)
an
d
h
y
b
r
id
m
o
d
els.
ViT
s
h
av
e
b
ee
n
s
h
o
wn
to
p
r
o
v
id
e
s
u
p
er
i
o
r
p
e
r
f
o
r
m
a
n
ce
in
h
ig
h
-
co
m
p
lex
ity
o
b
je
ct
d
etec
tio
n
task
s
,
an
d
th
ei
r
p
o
ten
tial
in
wee
d
d
etec
tio
n
ap
p
licatio
n
s
ca
n
b
e
f
u
r
th
er
ex
p
lo
r
e
d
to
o
v
er
c
o
m
e
th
e
lim
itatio
n
s
o
f
C
NN
-
b
ased
m
o
d
els
[
6
7
]
,
[
6
8
]
.
R
esear
ch
ca
n
also
co
n
s
id
er
a
h
y
b
r
id
a
p
p
r
o
ac
h
th
at
co
m
b
in
es
th
e
s
p
ee
d
o
f
YOL
Ov
5
with
th
e
s
eg
m
en
tatio
n
p
r
ec
is
io
n
o
f
Ma
s
k
R
-
C
NN
to
cr
ea
te
a
m
o
r
e
b
alan
ce
d
m
o
d
el.
Fo
r
ex
am
p
le,
a
co
m
b
in
atio
n
o
f
f
ea
tu
r
es
f
r
o
m
th
ese
two
m
o
d
els
co
u
ld
h
elp
a
d
d
r
ess
th
e
ch
allen
g
es
o
f
wee
d
d
etec
tio
n
in
ar
ea
s
with
t
o
p
o
g
r
ap
h
ic
v
ar
iatio
n
s
o
r
co
m
p
lex
cr
o
p
p
in
g
p
atter
n
s
.
S
u
ch
an
ap
p
r
o
ac
h
alig
n
s
with
th
e
r
ec
o
m
m
en
d
atio
n
s
,
wh
ich
s
u
g
g
est
ex
p
lo
r
in
g
in
n
o
v
ativ
e
tec
h
n
iq
u
es in
s
m
ar
t
ag
r
icu
ltu
r
e
t
o
im
p
r
o
v
e
ef
f
icien
cy
an
d
p
r
o
d
u
ctiv
ity
[
6
9
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Dro
n
e
-
a
s
s
is
ted
d
ee
p
lea
r
n
in
g
w
ee
d
d
etec
tio
n
fo
r
s
u
s
ta
in
a
b
le
a
g
r
icu
ltu
r
e
a
n
d
…
(
A
g
u
s
ta
n
La
tif
)
1437
4.
CO
NCLU
SI
O
N
T
h
is
r
esear
ch
s
h
o
ws
th
at
d
ee
p
lear
n
in
g
-
b
ased
wee
d
d
etec
t
io
n
m
o
d
els
o
f
f
er
s
ig
n
if
ican
t
ad
v
an
tag
es
o
v
er
tr
a
d
itio
n
al
m
eth
o
d
s
in
a
cc
u
r
ac
y
,
p
r
o
ce
s
s
in
g
tim
e
e
f
f
i
cien
cy
,
an
d
ad
a
p
tab
ilit
y
to
v
ar
io
u
s
ag
r
icu
ltu
r
al
en
v
ir
o
n
m
en
tal
co
n
d
itio
n
s
.
YOL
Ov
5
s
tan
d
s
o
u
t
with
a
h
ig
h
ac
cu
r
ac
y
o
f
9
4
%
an
d
a
p
r
o
ce
s
s
in
g
tim
e
o
f
o
n
ly
1
2
m
s
p
er
im
ag
e,
m
a
k
in
g
it
id
ea
l
f
o
r
r
ea
l
-
tim
e
ap
p
licatio
n
s
.
YOL
O
-
NAS
ac
h
iev
es
9
6
%
ac
cu
r
ac
y
,
s
u
itab
le
f
o
r
s
ce
n
ar
io
s
th
at
r
eq
u
ir
e
h
ig
h
p
r
ec
is
io
n
,
alb
eit
with
a
p
r
o
ce
s
s
in
g
tim
e
o
f
2
5
m
s
p
er
im
a
g
e.
Ma
s
k
R
-
C
NN
o
f
f
er
s
p
r
ec
is
e
s
eg
m
en
tatio
n
with
9
4
%
ac
cu
r
ac
y
,
b
u
t
is
lim
ited
b
y
a
3
1
m
s
p
er
-
im
a
g
e
p
r
o
ce
s
s
in
g
tim
e.
I
n
co
n
tr
ast,
th
e
tr
ad
itio
n
al
m
eth
o
d
s
h
o
ws
m
u
ch
lo
wer
p
er
f
o
r
m
an
ce
,
w
ith
an
av
er
ag
e
ac
c
u
r
ac
y
o
f
8
5
%
an
d
p
r
o
ce
s
s
in
g
tim
es
ex
ce
ed
in
g
5
0
m
s
,
lim
itin
g
its
s
ca
lab
ilit
y
f
o
r
p
r
ac
tical
u
s
e.
T
h
e
f
i
n
d
in
g
s
s
u
p
p
o
r
t
th
e
n
ee
d
f
o
r
AI
-
b
ased
tech
n
o
lo
g
ical
in
n
o
v
atio
n
s
to
o
v
er
co
m
e
co
n
s
tr
ain
ts
o
f
m
o
d
er
n
ag
r
ic
u
ltu
r
e,
esp
ec
ially
in
tr
o
p
ical
ar
ea
s
s
u
ch
as
Me
r
au
k
e.
Fo
r
f
u
tu
r
e
r
esear
ch
,
u
s
in
g
p
r
im
ar
y
d
atasets
an
d
ex
p
lo
r
in
g
m
eth
o
d
s
s
u
ch
as
YOL
Ov
8
,
YOL
O
-
NAS,
ViT
,
an
d
h
y
b
r
id
ap
p
r
o
ac
h
es c
an
b
e
s
tr
ateg
ic
s
tep
s
to
f
u
r
th
er
o
p
tim
ize
wee
d
d
etec
tio
n
p
er
f
o
r
m
an
ce
.
ACK
NO
WL
E
DG
M
E
N
T
S
T
h
e
a
u
t
h
o
r
s
g
r
ate
f
u
ll
y
ac
k
n
o
wle
d
g
e
t
h
e
Di
r
ec
t
o
r
at
e
o
f
R
esea
r
c
h
,
T
ec
h
n
o
l
o
g
y
,
a
n
d
C
o
m
m
u
n
it
y
Ser
v
i
ce
(
DR
T
PM
)
,
I
n
d
o
n
es
ia
,
f
o
r
s
u
p
p
o
r
ti
n
g
t
h
is
r
ese
a
r
c
h
.
T
h
e
a
u
t
h
o
r
s
a
ls
o
t
h
an
k
t
h
e
Fa
c
u
lty
o
f
E
n
g
i
n
e
e
r
i
n
g
,
Un
i
v
e
r
s
i
tas
N
e
g
e
r
i
Y
o
g
y
a
k
a
r
t
a
a
n
d
U
n
i
v
e
r
s
it
as
M
u
s
a
m
u
s
f
o
r
p
r
o
v
i
d
i
n
g
f
a
cili
ties
a
n
d
i
n
s
ti
tu
ti
o
n
al
s
u
p
p
o
r
t
d
u
r
i
n
g
t
h
e
r
esea
r
ch
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
is
r
esear
ch
was
f
u
n
d
ed
b
y
th
e
Dir
ec
to
r
ate
o
f
R
esear
ch
,
T
ec
h
n
o
lo
g
y
,
an
d
C
o
m
m
u
n
it
y
Ser
v
ice
(
DR
T
PM)
,
I
n
d
o
n
esia (
d
ec
r
ee
n
o
.
0
8
1
/C
3
/DT
.
0
5
.
0
0
/PL/2
0
2
5
;
co
n
tr
ac
t n
o
.
0
0
5
.
1
/UN.
5
2
.
8
/
L
T
/2
0
2
5
)
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Ag
u
s
tan
L
atif
✓
✓
✓
✓
✓
✓
✓
✓
✓
Han
d
ar
u
J
ati
✓
✓
✓
✓
✓
✓
Her
m
an
Dwi
Su
r
jo
n
o
✓
✓
✓
✓
Ma
n
i Y
u
s
u
f
✓
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
I
NF
O
RM
E
D
CO
NS
E
N
T
W
e
h
av
e
o
b
tain
ed
in
f
o
r
m
ed
c
o
n
s
en
t f
r
o
m
all
in
d
iv
id
u
als in
c
lu
d
ed
in
t
h
is
s
tu
d
y
.
E
T
H
I
CAL AP
P
RO
V
AL
No
t
ap
p
licab
le.
T
h
e
s
tu
d
y
is
b
ased
s
o
lely
o
n
ae
r
ial
im
a
g
er
y
ac
q
u
ir
ed
f
r
o
m
a
g
r
icu
ltu
r
al
f
ield
s
an
d
d
o
es n
o
t in
v
o
lv
e
h
u
m
an
p
ar
tic
ip
an
ts
,
an
im
als,
o
r
s
en
s
itiv
e
p
e
r
s
o
n
al
d
ata.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
p
r
im
a
r
y
UAV
d
ataset
co
llected
in
Me
r
a
u
k
e
(
1
,
3
1
8
im
a
g
es;
5
2
8
0
×
2
9
7
0
p
ix
els)
an
d
t
h
e
cu
r
ate
d
s
ec
o
n
d
ar
y
im
ag
e
d
ataset
(
4
,
0
5
0
im
ag
es;
5
2
8
0
×2
9
7
0
p
ix
els)
u
s
ed
f
o
r
m
o
d
el
tr
ain
i
n
g
an
d
ev
alu
atio
n
a
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.