I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Ro
bo
t
ics a
nd
Aut
o
m
a
t
io
n
(
I
J
RA)
Vo
l.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
,
p
p
.
52
~
62
I
SS
N:
2722
-
2
5
8
6
,
DOI
:
1
0
.
1
1
5
9
1
/i
jr
a
.
v
1
5
i
1
.
pp
52
-
62
52
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
r
a
.
ia
esco
r
e.
co
m
IST
D
-
LI
O
M
:
D
ir
ect
LiD
AR
-
i
nerti
a
l odo
metry a
nd
ma
pping
with
in
tensi
ty
-
en
ha
nced stable
t
ria
ng
le descrip
to
r
L
ix
ia
o
Ya
ng
1
,
Sh
eng
H
ua
1
,
Yo
ub
ing
F
eng
1
,
Sh
a
ng
zo
ng
Ya
ng
2
,
J
ie
Wa
ng
3
1
O
c
e
a
n
C
o
l
l
e
g
e
,
Ji
a
n
g
s
u
U
n
i
v
e
r
si
t
y
o
f
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
Z
h
e
n
j
i
a
n
g
,
C
h
i
n
a
2
C
o
m
p
u
t
e
r
C
o
l
l
e
g
e
,
J
i
a
n
g
s
u
U
n
i
v
e
r
si
t
y
o
f
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
Z
h
e
n
j
i
a
n
g
,
C
h
i
n
a
3
S
c
h
o
o
l
o
f
M
a
t
e
r
i
a
l
s
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
S
h
a
n
g
h
a
i
U
n
i
v
e
r
s
i
t
y
,
S
h
a
n
g
h
a
i
,
C
h
i
n
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l
12
,
2
0
2
5
R
ev
is
ed
No
v
21
,
2
0
2
5
Acc
ep
ted
Feb
9
,
2
0
2
6
To
a
d
d
re
ss
th
e
c
u
m
u
lati
v
e
d
rift
p
ro
b
lem
o
f
li
g
h
t
d
e
tec
ti
o
n
a
n
d
ra
n
g
i
n
g
(Li
DA
R)
-
in
e
rti
a
l
o
d
o
m
e
try
(LI
O)
in
lo
n
g
-
d
u
ra
ti
o
n
l
o
c
a
li
z
a
ti
o
n
a
n
d
m
a
p
p
in
g
tas
k
s,
th
is
p
a
p
e
r
p
ro
p
o
se
s
a
Li
D
AR
-
in
e
rti
a
l
o
d
o
m
e
try
a
n
d
m
a
p
p
i
n
g
sy
ste
m
,
in
ten
sit
y
-
e
n
h
a
n
c
e
d
sta
b
le
tri
a
n
g
l
e
d
e
sc
rip
to
r
-
L
iDAR
-
in
e
rti
a
l
o
d
o
m
e
try
a
n
d
m
a
p
p
in
g
(IS
TD
-
LIOM
)
,
b
a
se
d
o
n
t
h
e
in
ten
si
ty
-
e
n
h
a
n
c
e
d
sta
b
l
e
tri
a
n
g
le
d
e
sc
rip
to
r
(IS
TD).
T
h
is
s
y
ste
m
,
b
u
il
t
o
n
th
e
F
AST
-
LIO
2
fro
n
t
-
e
n
d
a
rc
h
it
e
c
tu
re
,
a
c
h
iev
e
s
g
l
o
b
a
l
c
o
n
siste
n
c
y
lo
c
a
li
z
a
ti
o
n
t
h
ro
u
g
h
l
o
o
p
c
lo
su
r
e
d
e
tec
ti
o
n
a
n
d
g
l
o
b
a
l
o
p
ti
m
iza
ti
o
n
.
F
irst,
we
d
e
sig
n
t
h
e
IS
TD
d
e
s
c
rip
to
r
b
y
c
o
m
b
in
i
n
g
g
e
o
m
e
tri
c
d
e
sc
rip
t
o
rs
o
f
tri
a
n
g
les
(in
c
lu
d
i
n
g
v
e
rtex
p
lan
e
n
o
rm
a
l
v
e
c
to
rs an
d
e
d
g
e
len
g
t
h
s) wit
h
lo
c
a
l
in
ten
sity
d
istri
b
u
ti
o
n
d
e
sc
rip
t
o
rs t
o
fo
r
m
a
c
o
m
p
a
c
t,
ro
tatio
n
-
i
n
v
a
ria
n
t
f
e
a
tu
re
re
p
re
se
n
tatio
n
.
Ne
x
t,
a
n
a
d
a
p
ti
v
e
k
e
y
fra
m
e
m
a
n
a
g
e
m
e
n
t
m
e
c
h
a
n
ism
is
c
o
n
stru
c
ted
,
w
h
ich
fi
lt
e
rs
k
e
y
fra
m
e
s
b
a
se
d
o
n
i
n
ter
-
fra
m
e
re
lativ
e
p
o
se
s
a
n
d
g
e
n
e
ra
tes
a
d
e
sc
rip
to
r
d
a
tab
a
se
.
A
h
y
b
rid
re
tr
iev
a
l
stra
teg
y
is
t
h
e
n
p
r
o
p
o
se
d
,
w
h
ich
c
o
m
b
in
e
s
d
e
sc
rip
to
r
sim
i
larity
m
a
tch
in
g
a
n
d
sp
a
ti
a
l
d
istan
c
e
fil
terin
g
,
f
o
rm
in
g
a
n
e
fficie
n
t
lo
o
p
c
lo
su
re
c
a
n
d
id
a
te
re
c
o
g
n
it
io
n
m
e
c
h
a
n
ism
.
Afte
r
a
p
p
ly
in
g
p
la
n
e
it
e
ra
ti
v
e
c
lo
se
st
p
o
i
n
t
(ICP
)
re
fi
n
e
m
e
n
t
a
n
d
g
e
o
m
e
tri
c
-
in
ten
sity
c
o
n
siste
n
c
y
v
a
li
d
a
ti
o
n
,
t
h
e
lo
o
p
c
lo
s
u
re
c
o
n
stra
in
ts
a
re
in
teg
ra
ted
in
t
o
a
p
o
se
g
ra
p
h
o
p
ti
m
iza
ti
o
n
fra
m
e
wo
rk
,
c
o
rre
c
t
in
g
o
d
o
m
e
try
d
rif
t.
Ex
p
e
rime
n
ts
o
n
th
e
KITT
I
d
a
tas
e
t
d
e
m
o
n
stra
te
th
a
t
th
e
I
S
TD
-
LIOM
s
y
ste
m
sig
n
ifi
c
a
n
tl
y
e
n
h
a
n
c
e
s
m
a
p
g
lo
b
a
l
c
o
n
siste
n
c
y
wh
il
e
m
a
in
tain
i
n
g
re
a
l
-
ti
m
e
c
o
m
p
u
tati
o
n
a
l
p
e
rfo
rm
a
n
c
e
.
K
ey
w
o
r
d
s
:
L
I
DAR in
ten
s
ity
L
o
o
p
d
etec
tio
n
Sen
s
o
r
f
u
s
io
n
Simu
ltan
eo
u
s
lo
ca
lizatio
n
a
n
d
m
ap
p
in
g
T
r
ian
g
u
lar
d
escr
ip
to
r
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
:
Yo
u
b
in
g
Fen
g
Oce
an
C
o
lleg
e,
J
ian
g
s
u
Un
iv
e
r
s
ity
o
f
Scien
ce
an
d
T
ec
h
n
o
lo
g
y
2
Me
n
g
x
i Ro
ad
,
J
in
g
k
o
u
Dis
tr
ict,
Z
h
en
jian
g
,
J
ian
g
s
u
2
1
2
0
0
0
,
C
h
in
a
E
m
ail:
y
zf
y
b
@
ju
s
t.e
d
u
.
cn
1.
I
NT
RO
D
UCT
I
O
N
L
ig
h
t
d
etec
tio
n
an
d
r
an
g
in
g
(
L
iDAR
)
-
in
er
tial
o
d
o
m
etr
y
(
L
I
O)
[
1
]
–
[
4
]
an
d
L
iDAR
-
in
er
tial
o
d
o
m
etr
y
an
d
m
ap
p
i
n
g
(
L
I
OM
)
[
5
]
–
[
8
]
h
av
e
g
ain
ed
s
ig
n
if
ica
n
t
p
r
o
m
in
en
ce
in
th
e
f
ield
s
o
f
r
o
b
o
tic
n
av
ig
atio
n
an
d
au
to
n
o
m
o
u
s
d
r
iv
in
g
in
r
ec
en
t
y
ea
r
s
.
T
h
ese
tech
n
o
lo
g
ies
en
a
b
le
r
ea
l
-
tim
e
en
v
ir
o
n
m
en
tal
p
er
ce
p
tio
n
an
d
m
ap
co
n
s
tr
u
ctio
n
b
y
f
u
s
in
g
d
ata
f
r
o
m
lig
h
t
d
etec
tio
n
a
n
d
r
a
n
g
in
g
(
L
iDAR
)
an
d
i
n
er
tial
m
ea
s
u
r
em
en
t
u
n
its
(
I
MU
)
.
T
h
is
m
u
lti
-
s
en
s
o
r
f
u
s
io
n
a
p
p
r
o
ac
h
n
o
t
o
n
ly
en
h
a
n
ce
s
th
e
a
cc
u
r
ac
y
an
d
r
o
b
u
s
tn
ess
o
f
lo
ca
lizatio
n
b
u
t
also
p
r
o
v
id
es a
f
o
u
n
d
atio
n
al
f
r
am
e
wo
r
k
f
o
r
n
a
v
ig
atio
n
.
T
h
e
in
cr
e
asin
g
ap
p
licatio
n
o
f
L
I
O
an
d
L
I
OM
tech
n
o
lo
g
ies
in
au
to
n
o
m
o
u
s
d
r
iv
in
g
[
9
]
,
[
1
0
]
,
u
n
m
an
n
e
d
ae
r
ial
v
eh
icle
s
(
UAVs)
[
1
1
]
,
[
1
2
]
,
a
n
d
r
o
b
o
tics
h
ig
h
lig
h
t
th
eir
g
r
o
win
g
im
p
o
r
ta
n
ce
,
o
f
f
er
in
g
p
o
wer
f
u
l
tech
n
ical
s
u
p
p
o
r
t
f
o
r
ac
h
iev
in
g
au
to
n
o
m
o
u
s
n
av
ig
atio
n
a
n
d
en
v
ir
o
n
m
en
tal
p
er
ce
p
tio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
I
S
TD
-
LIO
M
:
Dir
ec
t
LiDAR
-
in
erti
a
l o
d
o
metry
a
n
d
ma
p
p
in
g
w
ith
in
ten
s
ity
-
en
h
a
n
ce
d
s
ta
b
le
…
(
Lixia
o
Ya
n
g
)
53
L
I
O
o
f
f
e
r
s
a
s
ig
n
if
ica
n
t
ad
v
a
n
tag
e
f
o
r
a
u
to
n
o
m
o
u
s
r
o
b
o
ts
[
1
3
]
o
p
e
r
atin
g
i
n
d
y
n
am
ic
en
v
ir
o
n
m
e
n
ts
b
y
co
m
b
i
n
in
g
L
iDAR
's
p
r
ec
is
e
g
eo
m
etr
ic
s
en
s
in
g
with
th
e
h
ig
h
-
f
r
e
q
u
en
c
y
m
o
tio
n
tr
ac
k
in
g
p
r
o
v
id
ed
b
y
I
MU
s
.
Ho
wev
er
,
d
r
if
t
er
r
o
r
s
s
till
em
er
g
e
o
v
er
ex
ten
d
e
d
p
er
i
o
d
s
o
f
o
p
e
r
atio
n
.
T
h
ese
er
r
o
r
s
a
r
is
e
d
u
e
to
f
ac
to
r
s
s
u
ch
as
s
en
s
o
r
n
o
is
e,
e
n
v
ir
o
n
m
en
tal
ch
an
g
es,
an
d
th
e
in
h
e
r
en
t
ac
cu
m
u
latio
n
o
f
er
r
o
r
s
w
ith
in
th
e
al
g
o
r
ith
m
its
elf
.
C
u
m
u
lativ
e
d
r
if
t
ca
n
d
eg
r
ad
e
th
e
r
o
b
o
t'
s
lo
ca
lizatio
n
ac
cu
r
ac
y
a
n
d
m
ap
q
u
ality
,
th
u
s
lim
itin
g
its
p
er
f
o
r
m
an
ce
in
lo
n
g
-
d
is
tan
ce
n
av
ig
atio
n
a
n
d
co
m
p
lex
en
v
ir
o
n
m
en
ts
.
So
m
e
s
tu
d
ies
h
av
e
a
d
d
r
ess
ed
th
is
is
s
u
e
b
y
u
s
in
g
l
o
o
p
clo
s
u
r
e
d
etec
ti
o
n
to
id
e
n
tify
wh
en
a
r
o
b
o
t
r
ev
is
its
a
p
r
ev
io
u
s
ly
en
c
o
u
n
te
r
ed
lo
ca
tio
n
,
t
h
er
eb
y
elim
in
atin
g
cu
m
u
lativ
e
er
r
o
r
s
th
r
o
u
g
h
r
eg
is
tr
atio
n
an
d
im
p
r
o
v
i
n
g
th
e
co
n
s
is
ten
cy
o
f
lo
ca
lizatio
n
an
d
m
ap
p
in
g
.
E
ar
ly
l
o
o
p
cl
o
s
u
r
e
m
eth
o
d
s
p
r
im
a
r
ily
r
elied
o
n
d
is
tan
ce
m
etr
ics,
co
m
p
ar
in
g
th
e
cu
r
r
en
t
m
ap
with
h
is
to
r
ical
m
ap
s
to
d
etec
t
lo
o
p
s
.
W
h
ile
s
im
p
le
an
d
in
tu
itiv
e,
th
is
ap
p
r
o
ac
h
is
p
r
o
n
e
to
f
alse
p
o
s
itiv
es
an
d
m
is
s
ed
d
etec
tio
n
s
in
co
m
p
lex
en
v
ir
o
n
m
en
ts
.
T
o
im
p
r
o
v
e
th
e
r
ec
all
r
ate
o
f
lo
o
p
clo
s
u
r
e
d
et
ec
tio
n
,
r
esear
ch
er
s
h
av
e
p
r
o
p
o
s
ed
v
ar
io
u
s
d
escr
ip
to
r
s
to
ex
tr
ac
t
s
p
atial
s
tr
u
ctu
r
al
f
ea
tu
r
es
f
r
o
m
L
iDAR
p
o
in
t
clo
u
d
s
.
B
y
r
etr
iev
in
g
s
im
ilar
d
escr
ip
to
r
s
f
r
o
m
a
d
atab
ase,
lo
o
p
clo
s
u
r
e
f
r
am
es
ca
n
b
e
id
en
tifie
d
with
lo
w
co
m
p
u
tatio
n
al
co
s
t.
Scan
c
o
n
tex
t
[
1
4
]
is
a
p
o
in
t
cl
o
u
d
d
escr
ip
tio
n
alg
o
r
ith
m
wh
o
s
e
co
r
e
id
ea
is
to
p
r
o
ject
3
D
p
o
in
t
clo
u
d
d
ata
in
to
a
2
D
p
o
lar
co
o
r
d
i
n
ate
s
p
ac
e,
f
o
r
m
in
g
a
r
in
g
-
s
ec
to
r
m
atr
ix
.
Du
e
to
th
e
u
s
e
o
f
p
o
lar
co
o
r
d
in
ates,
s
im
ilar
s
ce
n
es
ca
n
s
till
b
e
m
atch
ed
b
y
h
o
r
izo
n
tally
s
h
if
ti
n
g
th
e
m
atr
ix
,
ev
en
wh
e
n
th
e
p
o
in
t
clo
u
d
d
ata
co
n
tain
s
an
g
u
lar
d
ev
iatio
n
s
.
Scan
c
o
n
tex
t++
[
1
5
]
is
an
im
p
r
o
v
ed
v
er
s
io
n
o
f
s
ca
n
c
o
n
tex
t
,
wh
ich
en
h
an
ce
s
th
e
d
is
cr
im
in
ativ
e
p
o
we
r
an
d
r
o
b
u
s
tn
ess
o
f
th
e
d
escr
ip
to
r
b
y
i
n
tr
o
d
u
ci
n
g
m
o
r
e
s
o
p
h
is
ticated
f
ea
tu
r
e
en
c
o
d
in
g
tech
n
iq
u
es
.
B
in
ar
y
tr
ee
c
o
d
e
(
B
T
C
)
[
1
6
]
is
an
al
g
o
r
ith
m
d
esig
n
ed
f
o
r
en
v
i
r
o
n
m
e
n
t
r
e
co
g
n
itio
n
an
d
lo
o
p
clo
s
u
r
e
d
etec
tio
n
u
s
in
g
L
iDAR
p
o
in
t
clo
u
d
d
ata.
I
t
co
n
s
tr
u
cts
a
b
in
ar
y
tr
ee
s
tr
u
ctu
r
e
to
r
ep
r
esen
t
p
o
in
t
clo
u
d
d
ata
ef
f
icien
tly
an
d
m
atch
it,
en
ab
lin
g
f
ast
en
v
ir
o
n
m
en
t
r
ec
o
g
n
itio
n
a
n
d
lo
o
p
d
etec
tio
n
.
R
ef
er
en
ce
[
1
7
]
p
r
o
p
o
s
es
a
g
lo
b
al
d
escr
ip
to
r
f
o
r
3
D
p
lace
r
ec
o
g
n
itio
n
,
wh
er
e
th
e
c
o
r
e
id
ea
is
t
o
d
escr
ib
e
lo
ca
l
k
ey
p
o
in
ts
in
th
e
3
D
p
o
in
t
clo
u
d
u
s
in
g
t
h
e
ed
g
e
len
g
t
h
s
an
d
an
g
les
o
f
t
r
ian
g
les,
th
er
e
b
y
ac
h
ie
v
in
g
e
f
f
icien
t
p
lace
r
ec
o
g
n
itio
n
an
d
l
o
o
p
cl
o
s
u
r
e
d
etec
tio
n
.
Alth
o
u
g
h
ex
is
tin
g
d
escr
ip
to
r
-
b
ased
m
eth
o
d
s
ca
n
p
e
r
f
o
r
m
r
ap
id
lo
o
p
clo
s
u
r
e
d
etec
tio
n
[
1
8
]
–
[
2
0
]
,
th
ey
s
till
s
u
f
f
er
f
r
o
m
m
is
s
ed
d
etec
tio
n
s
an
d
f
alse
p
o
s
itiv
es
in
co
m
p
lex
e
n
v
ir
o
n
m
en
ts
.
T
o
f
u
r
th
er
en
h
an
ce
lo
o
p
clo
s
u
r
e
d
etec
tio
n
p
e
r
f
o
r
m
an
ce
,
th
is
p
ap
er
p
r
o
p
o
s
es
s
ev
er
al
i
m
p
r
o
v
e
m
en
ts
.
W
e
d
esig
n
th
e
in
ten
s
ity
-
en
h
an
ce
d
s
tab
le
tr
ian
g
le
d
escr
ip
to
r
(
I
ST
D)
,
wh
ich
in
n
o
v
ativ
ely
in
teg
r
ates
a
tr
ian
g
u
lar
g
eo
m
etr
ic
d
e
s
cr
ip
to
r
—
b
ased
o
n
v
er
tex
p
lan
e
n
o
r
m
al
v
ec
t
o
r
s
an
d
ed
g
e
len
g
th
f
ea
tu
r
es
—
with
a
lo
ca
l
in
ten
s
ity
d
is
tr
ib
u
tio
n
d
escr
ip
to
r
to
co
n
s
tr
u
ct
a
co
m
p
ac
t,
r
o
tatio
n
-
in
v
ar
ian
t
f
ea
tu
r
e
r
ep
r
esen
tati
o
n
.
T
h
is
m
u
ltimo
d
al
f
u
s
ed
d
escr
ip
to
r
en
h
a
n
ce
s
tr
ad
itio
n
al
g
eo
m
etr
ic
s
tr
u
ct
u
r
e
s
b
y
in
co
r
p
o
r
atin
g
i
n
ten
s
ity
in
f
o
r
m
atio
n
,
th
er
eb
y
s
ig
n
if
ican
tl
y
im
p
r
o
v
in
g
s
ce
n
e
d
is
cr
im
in
ab
ilit
y
.
Fu
r
th
e
r
m
o
r
e
,
th
is
p
ap
e
r
p
r
o
p
o
s
es
th
e
L
iDAR
SLAM
f
r
am
ewo
r
k
I
STD
-
L
I
OM
,
wh
ic
h
in
teg
r
ates
d
y
n
am
ic
k
ey
f
r
a
m
e
m
an
ag
em
en
t,
d
escr
ip
to
r
m
atch
in
g
,
an
d
d
is
tan
ce
-
co
n
s
tr
ain
ed
lo
o
p
cl
o
s
u
r
e
d
etec
tio
n
to
b
u
ild
a
g
lo
b
ally
co
n
s
is
ten
t
b
ac
k
-
en
d
o
p
tim
izat
io
n
s
y
s
tem
.
B
ased
o
n
th
e
FA
ST
-
L
I
O2
[
2
1
]
f
r
o
n
t
en
d
,
th
e
f
r
am
ewo
r
k
ex
tr
ac
ts
d
escr
ip
to
r
s
f
r
o
m
ea
ch
k
ey
f
r
a
m
e
b
y
f
u
s
in
g
in
ten
s
ity
an
d
g
eo
m
etr
ic
f
ea
tu
r
es.
I
t
th
en
p
er
f
o
r
m
s
lo
o
p
clo
s
u
r
e
ca
n
d
id
ate
id
en
tific
atio
n
th
r
o
u
g
h
d
escr
ip
t
o
r
s
im
ilar
ity
m
a
tch
in
g
an
d
s
p
atial
p
r
o
x
im
ity
c
o
n
s
tr
ain
ts
,
ef
f
ec
tiv
ely
s
u
p
p
r
ess
in
g
o
d
o
m
etr
y
d
r
if
t
o
v
er
lo
n
g
-
ter
m
s
eq
u
en
ce
s
an
d
g
r
ea
tly
en
h
an
cin
g
o
v
er
all
s
y
s
tem
co
n
s
is
ten
cy
.
E
x
p
er
im
en
tal
r
esu
lts
o
n
th
e
KI
T
T
I
0
5
/0
7
/0
8
d
atasets
[
2
2
]
,
[
2
3
]
d
em
o
n
s
tr
ate
th
at
th
e
I
STD
-
L
I
OM
s
y
s
tem
,
t
h
r
o
u
g
h
h
ig
h
-
p
r
e
cisi
o
n
lo
o
p
cl
o
s
u
r
e
d
etec
tio
n
an
d
f
ac
t
o
r
g
r
ap
h
o
p
tim
izatio
n
,
s
ig
n
if
ican
tly
im
p
r
o
v
es
lo
ca
liz
atio
n
ac
cu
r
ac
y
an
d
e
n
h
an
ce
s
th
e
g
lo
b
al
co
n
s
is
ten
cy
o
f
lar
g
e
-
s
ca
le
p
o
in
t
clo
u
d
m
ap
s
,
p
r
o
v
id
in
g
r
eliab
le
tech
n
ical
s
u
p
p
o
r
t f
o
r
au
t
o
n
o
m
o
u
s
n
av
ig
atio
n
in
c
o
m
p
lex
en
v
ir
o
n
m
en
ts
.
T
h
e
r
em
ain
d
er
o
f
th
is
p
ap
er
is
o
r
g
an
ize
d
as
f
o
llo
ws.
Sectio
n
I
I
in
tr
o
d
u
ce
s
an
d
an
aly
ze
s
th
e
p
r
o
p
o
s
ed
m
eth
o
d
.
Sectio
n
I
I
I
p
r
esen
ts
th
e
ex
p
er
im
e
n
tal
r
esu
lts
o
n
th
e
KI
T
T
I
d
ataset,
an
d
Sectio
n
I
V
co
n
clu
d
es
th
is
wo
r
k
.
2.
M
E
T
H
O
DO
L
O
G
Y
2
.
1
.
F
r
a
m
ewo
r
k
o
v
er
v
iew
As
s
h
o
wn
in
Fig
u
r
e
1
,
th
e
s
y
s
tem
f
r
am
ewo
r
k
p
r
esen
te
d
in
th
is
p
ap
er
is
a
co
m
p
lete
SLAM
s
y
s
tem
th
at
in
clu
d
es
b
o
th
a
f
r
o
n
t
-
en
d
L
I
O
a
n
d
a
b
ac
k
-
en
d
f
a
cto
r
g
r
ap
h
o
p
tim
izatio
n
.
T
h
e
f
r
o
n
t
en
d
is
b
ased
o
n
th
e
FAST
-
L
I
O2
ar
ch
itectu
r
e,
u
til
izin
g
an
iter
ativ
e
Kalm
an
f
il
ter
-
b
ased
s
tate
esti
m
atio
n
m
eth
o
d
to
ac
h
iev
e
a
tig
h
tly
co
u
p
led
f
u
s
io
n
o
f
L
iDAR
an
d
I
MU
d
ata.
Af
ter
p
r
o
c
ess
in
g
ea
ch
L
iDAR
f
r
am
e,
th
e
f
r
o
n
t
-
en
d
m
o
d
u
le
tr
an
s
m
its
th
e
u
n
d
is
to
r
ted
p
o
in
t
clo
u
d
an
d
its
co
r
r
esp
o
n
d
in
g
p
o
s
e
esti
m
ate
to
th
e
b
ac
k
en
d
f
o
r
l
o
o
p
clo
s
u
r
e
d
etec
tio
n
an
d
g
lo
b
al
o
p
tim
izatio
n
.
T
h
e
b
ac
k
-
en
d
m
o
d
u
le
a
d
ap
tiv
ely
s
elec
ts
k
ey
f
r
am
es
b
ased
o
n
th
e
r
elativ
e
p
o
s
e
ch
an
g
es
an
d
tim
e
in
ter
v
als.
Fo
r
ea
ch
n
ewly
s
elec
ted
k
ey
f
r
am
e
,
th
e
s
y
s
tem
ac
cu
m
u
lates
p
o
in
t
cl
o
u
d
s
f
r
o
m
all
f
r
am
es
b
etwe
en
t
h
e
cu
r
r
en
t
an
d
p
r
e
v
io
u
s
k
ey
f
r
am
es,
f
o
llo
wed
b
y
v
o
x
el
-
b
ased
d
o
wn
s
am
p
lin
g
t
o
f
o
r
m
an
e
n
h
an
ce
d
p
o
i
n
t
clo
u
d
with
r
ich
er
s
p
atial
in
f
o
r
m
atio
n
.
T
h
is
ac
cu
m
u
latio
n
s
tr
ateg
y
e
f
f
ec
tiv
ely
ex
p
a
n
d
s
th
e
s
p
atial
co
v
er
ag
e
o
f
a
s
in
g
le
f
r
am
e
p
o
in
t
clo
u
d
,
p
r
o
v
id
in
g
a
m
o
r
e
co
m
p
r
eh
e
n
s
iv
e
en
v
ir
o
n
m
en
ta
l
r
ep
r
esen
tatio
n
f
o
r
s
u
b
s
eq
u
en
t
d
escr
ip
to
r
g
e
n
er
atio
n
an
d
l
o
o
p
clo
s
u
r
e
d
etec
tio
n
.
I
t
also
s
ig
n
if
ican
tly
r
e
d
u
ce
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
52
-
62
54
th
e
p
r
o
b
a
b
ilit
y
o
f
m
is
m
atch
es
ca
u
s
ed
b
y
th
e
v
iewp
o
in
t
lim
itatio
n
s
o
f
in
d
iv
id
u
al
p
o
in
t
clo
u
d
s
.
B
ased
o
n
th
e
en
h
an
ce
d
p
o
in
t
clo
u
d
,
th
e
b
ac
k
-
en
d
m
o
d
u
le
e
x
tr
ac
ts
I
STD
d
escr
ip
to
r
s
.
I
t
b
u
ild
s
a
d
escr
i
p
to
r
d
atab
ase
wh
ile
also
r
ec
o
r
d
in
g
th
e
r
elativ
e
p
o
s
e
co
n
s
tr
ain
ts
b
etwe
en
k
ey
f
r
a
m
es
as
th
e
in
itial
ed
g
es
in
th
e
f
ac
to
r
g
r
ap
h
.
T
h
e
lo
o
p
clo
s
u
r
e
d
etec
tio
n
m
o
d
u
le
id
en
tifie
s
p
o
ten
tial
lo
o
p
cl
o
s
u
r
e
ca
n
d
id
ates
th
r
o
u
g
h
a
d
u
al
-
p
ath
ca
n
d
id
ate
f
r
am
e
s
elec
tio
n
m
ec
h
an
is
m
,
w
h
ich
co
m
b
in
es
d
escr
ip
t
o
r
s
im
ilar
ity
m
atch
in
g
with
s
p
atial
d
is
tan
ce
co
n
s
tr
ain
ts
.
T
h
ese
ca
n
d
id
ates
ar
e
th
e
n
v
e
r
if
ied
f
o
r
g
eo
m
etr
ic
co
n
s
is
ten
cy
u
s
in
g
a
p
o
in
t
-
to
-
p
lan
e
iter
ativ
e
clo
s
est
p
o
in
t
(
I
C
P
)
r
eg
is
tr
atio
n
alg
o
r
ith
m
.
L
o
o
p
clo
s
u
r
e
co
n
s
tr
ain
ts
th
at
m
ee
t
th
e
g
eo
m
etr
ic
co
n
s
is
ten
cy
cr
iter
ia
ar
e
ad
d
ed
to
th
e
f
ac
to
r
g
r
a
p
h
,
a
n
d
th
e
g
l
o
b
al
p
o
s
es
ar
e
c
o
r
r
ec
ted
th
r
o
u
g
h
n
o
n
lin
ea
r
o
p
tim
izatio
n
.
Fin
ally
,
th
e
s
y
s
tem
r
e
-
r
eg
is
ter
s
th
e
k
ey
f
r
am
e
p
o
in
t
c
lo
u
d
s
u
s
in
g
th
e
o
p
tim
ized
p
o
s
e
tr
ajec
to
r
y
,
g
en
er
atin
g
a
g
lo
b
ally
co
n
s
is
ten
t
3
D
L
I
DAR p
o
in
ts
m
ap
.
Fig
u
r
e
1
.
Sy
s
tem
o
v
er
v
iew
o
f
I
STD
2
.
2
.
S
ta
b
le tria
n
g
le d
e
sc
r
ip
to
r
I
n
s
p
ir
ed
b
y
[
2
4
]
,
to
im
p
r
o
v
e
s
eg
m
en
tatio
n
s
tab
ilit
y
,
lo
o
p
clo
s
u
r
e
d
etec
tio
n
is
p
er
f
o
r
m
ed
o
n
k
ey
f
r
am
es,
wh
ic
h
co
n
s
is
t
o
f
p
o
in
ts
ac
cu
m
u
lated
f
r
o
m
s
ev
e
r
al
co
n
s
ec
u
tiv
e
s
ca
n
s
,
th
u
s
r
e
s
u
ltin
g
in
in
cr
ea
s
ed
p
o
in
t
clo
u
d
d
e
n
s
ity
r
eg
ar
d
les
s
o
f
th
e
s
p
ec
if
ic
L
iDA
R
s
ca
n
n
in
g
p
atter
n
.
Sp
ec
if
ically
,
we
u
tili
ze
L
iDA
R
o
d
o
m
etr
y
[
2
5
]
to
r
eg
is
ter
ea
c
h
n
ew
in
p
u
t
p
o
in
t
clo
u
d
with
th
e
cu
r
r
en
t
k
e
y
f
r
am
e
.
New
k
ey
f
r
am
es
ar
e
cr
ea
ted
wh
en
th
e
n
u
m
b
er
o
f
s
u
b
f
r
am
e
s
ac
cu
m
u
lates to
a
ce
r
tain
th
r
e
s
h
o
ld
.
2
.
2
.
1
.
L
I
DAR
p
la
ne
a
nd
k
ey
po
ints det
ec
t
io
n
Fo
r
a
g
iv
en
k
e
y
f
r
am
e
o
f
th
e
p
o
in
t
clo
u
d
,
p
lan
e
d
etec
tio
n
is
i
n
itially
p
er
f
o
r
m
ed
u
s
in
g
r
e
g
io
n
g
r
o
win
g
.
Sp
ec
if
ically
,
th
e
en
tire
p
o
in
t
c
lo
u
d
is
d
iv
id
ed
in
to
v
o
x
els
o
f
a
g
iv
en
s
ize
(
e.
g
.
,
1
m
)
.
E
ac
h
v
o
x
el
co
n
tain
s
a
s
et
o
f
p
o
i
n
ts
(
=
1
,
.
.
.
.
,
)
,
an
d
th
e
co
v
a
r
ian
ce
m
atr
ix
o
f
t
h
e
p
o
i
n
ts
is
co
m
p
u
t
ed
:
=
1
∑
;
∑
=
1
∑
(
−
)
=
1
=
1
(
−
)
(
1
)
L
et
r
ep
r
esen
t th
e
k
-
th
lar
g
est eig
en
v
alu
e
o
f
m
atr
ix
∑
.
T
h
e
p
la
n
e
cr
iter
io
n
is
d
e
f
in
ed
as:
3
<
1
2
>
2
wh
e
r
e
t
h
e
d
e
f
au
lt
v
al
u
es f
o
r
1
a
n
d
2
ar
e
0
.
0
1
a
n
d
0
.
0
5
,
r
es
p
e
cti
v
el
y
.
Usi
n
g
t
h
is
c
r
it
er
i
o
n
,
we
c
an
d
ete
r
m
in
e
i
f
th
e
p
o
i
n
ts
wit
h
i
n
th
e
v
o
x
el
f
o
r
m
a
p
la
n
e
.
I
f
s
o
,
t
h
e
v
o
x
el
is
l
ab
ele
d
as
a
p
la
n
e
v
o
x
el
.
S
ta
r
ti
n
g
f
r
o
m
a
n
y
p
la
n
e
v
o
x
el
,
a
p
la
n
e
is
i
n
it
iali
ze
d
,
a
n
d
t
h
e
p
la
n
e
is
e
x
p
a
n
d
e
d
b
y
s
ea
r
c
h
i
n
g
f
o
r
a
d
ja
ce
n
t
v
o
x
e
ls
.
I
f
an
a
d
j
ac
en
t
v
o
x
el
b
el
o
n
g
s
to
t
h
e
s
am
e
p
l
an
e
(
w
it
h
th
e
s
a
m
e
p
la
n
e'
s
n
o
r
m
a
l
d
i
r
e
cti
o
n
an
d
d
is
ta
n
ce
b
el
o
w
a
t
h
r
esh
o
l
d
)
,
it
is
a
d
d
e
d
to
t
h
e
g
r
o
wi
n
g
p
l
a
n
e
.
Ot
h
e
r
w
is
e,
i
f
t
h
e
a
d
ja
ce
n
t
v
o
x
el
d
o
es
n
o
t
b
el
o
n
g
to
t
h
e
s
a
m
e
p
la
n
e
,
it
is
ad
d
ed
t
o
t
h
e
b
o
u
n
d
a
r
y
v
o
x
el
lis
t
o
f
t
h
e
g
r
o
win
g
p
la
n
e
.
T
h
is
p
r
o
ce
s
s
is
r
e
p
e
ate
d
u
n
til
a
ll
a
d
j
ac
en
t
v
o
x
el
s
h
a
v
e
b
e
en
ad
d
ed
o
r
th
e
b
o
u
n
d
ar
y
v
o
x
els
a
r
e
r
e
a
ch
ed
.
F
o
r
b
o
u
n
d
a
r
y
v
o
x
e
ls
,
th
e
p
o
i
n
ts
t
h
e
y
c
o
n
t
ai
n
a
r
e
p
r
o
j
ec
te
d
o
n
to
t
h
e
co
r
r
es
p
o
n
d
i
n
g
p
l
an
e.
F
o
r
ea
c
h
p
l
a
n
e
,
w
e
cr
ea
t
e
an
i
m
a
g
e
w
h
e
r
e
t
h
e
im
a
g
e
p
l
a
n
e
c
o
i
n
ci
d
e
s
wit
h
th
e
d
e
te
cte
d
p
la
n
e,
a
n
d
e
ac
h
p
i
x
el
r
e
p
r
es
en
ts
t
h
e
m
ax
im
u
m
d
is
t
an
ce
o
f
p
o
i
n
ts
c
o
n
tai
n
ed
wit
h
i
n
t
h
e
b
o
u
n
d
a
r
y
v
o
x
e
ls
o
f
t
h
e
p
la
n
e.
A
k
e
y
p
o
i
n
t
is
th
en
s
e
le
cte
d
as
th
e
p
o
i
n
t
wi
th
t
h
e
l
ar
g
e
s
t
p
i
x
el
v
al
u
e
w
it
h
i
n
i
ts
5
×
5
n
e
ig
h
b
o
r
h
o
o
d
.
2
.
2
.
2
.
T
ria
ng
ula
r
des
cr
ipto
r
co
ns
t
ruct
io
n
Usi
n
g
th
e
k
e
y
p
o
in
ts
ex
tr
ac
t
ed
f
r
o
m
th
e
k
e
y
f
r
am
es,
a
k
-
D
tr
ee
is
co
n
s
tr
u
cte
d
,
a
n
d
2
0
n
ea
r
est
n
eig
h
b
o
r
s
ar
e
s
ea
r
ch
ed
f
o
r
ea
ch
p
o
in
t
to
f
o
r
m
tr
ian
g
u
lar
d
escr
ip
to
r
s
.
R
ed
u
n
d
an
t
d
escr
ip
to
r
s
with
id
en
tical
ed
g
e
len
g
th
s
ar
e
r
em
o
v
ed
.
E
ac
h
tr
ian
g
u
lar
d
escr
ip
t
o
r
co
n
s
is
ts
o
f
th
r
ee
v
er
tices
1
,
2
,
an
d
3
,
alo
n
g
with
th
eir
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
I
S
TD
-
LIO
M
:
Dir
ec
t
LiDAR
-
in
erti
a
l o
d
o
metry
a
n
d
ma
p
p
in
g
w
ith
in
ten
s
ity
-
en
h
a
n
ce
d
s
ta
b
le
…
(
Lixia
o
Ya
n
g
)
55
co
r
r
esp
o
n
d
in
g
p
r
o
jecte
d
n
o
r
m
al
v
ec
to
r
s
1
,
2
an
d
3
.
Ad
d
itio
n
ally
,
th
e
v
er
tices
o
f
th
e
tr
ian
g
le
ar
e
ar
r
an
g
e
d
in
d
escen
d
in
g
o
r
d
er
o
f
ed
g
e
len
g
th
s
(
s
ee
Fig
u
r
e
2
)
.
T
h
is
ed
g
e
len
g
th
s
o
r
tin
g
en
s
u
r
es th
at
12
≥
23
≥
13
.
Fig
u
r
e
2
.
T
r
ian
g
u
la
r
d
escr
ip
to
r
2
.
3
.
L
o
ca
l
inte
ns
it
y
f
e
a
t
ure
ex
t
ra
ct
io
n
Alth
o
u
g
h
tr
ian
g
u
lar
d
escr
ip
to
r
s
ar
e
ca
p
ab
le
o
f
ca
p
tu
r
in
g
s
p
atial
s
tr
u
ctu
r
al
f
ea
tu
r
es,
th
ey
ar
e
p
r
o
n
e
to
m
is
m
atch
es
in
r
eg
io
n
s
with
g
e
o
m
etr
ically
s
im
ilar
s
tr
u
ctu
r
es.
L
iDAR
in
ten
s
ity
in
f
o
r
m
atio
n
,
wh
ich
r
ef
lects
th
e
s
u
r
f
ac
e
r
ef
lecta
n
ce
ch
ar
ac
ter
is
tics
o
f
o
b
jects,
ca
n
d
is
tin
g
u
is
h
b
etwe
en
d
if
f
er
en
t
m
ater
ials
.
T
h
er
ef
o
r
e,
in
ten
s
ity
in
f
o
r
m
atio
n
is
in
co
r
p
o
r
ated
t
o
en
h
an
ce
th
e
d
is
cr
im
in
ativ
e
p
o
wer
o
f
th
e
d
escr
ip
to
r
s
in
s
im
il
ar
s
ce
n
es.
Fo
r
ea
ch
v
er
tex
o
f
th
e
tr
ian
g
u
lar
d
escr
ip
to
r
,
a
lo
ca
l
in
ten
s
ity
d
escr
ip
to
r
is
co
n
s
tr
u
cted
.
I
n
t
h
e
d
ir
ec
tio
n
o
f
th
e
p
lan
e
co
n
tain
in
g
th
e
v
er
tex
,
th
e
in
ten
s
ity
d
is
tr
ib
u
tio
n
ac
r
o
s
s
d
if
f
e
r
en
t
h
eig
h
t
lev
els
r
e
f
lects
th
e
m
ater
ial
ch
ar
ac
ter
is
tics
o
f
th
e
f
ea
tu
r
e
in
th
e
v
er
tical
d
ir
ec
tio
n
,
th
e
r
eb
y
im
p
r
o
v
in
g
its
d
is
tin
ctiv
en
ess
.
Sp
ec
if
ically
,
d
if
f
er
en
t
o
b
jects
s
u
ch
as
b
u
il
d
in
g
s
,
v
eg
etatio
n
,
an
d
o
th
er
p
r
o
tr
u
s
io
n
s
ex
h
ib
it
d
is
tin
ct
in
ten
s
ity
d
is
tr
ib
u
tio
n
s
alo
n
g
th
e
h
eig
h
t
d
ir
ec
tio
n
.
2
.
3
.
1
.
Ve
rt
ic
a
l
s
a
m
p
lin
g
s
t
r
a
t
eg
y
des
i
g
n
Giv
en
a
k
ey
p
o
in
t
an
d
th
e
u
n
i
t
n
o
r
m
al
v
ec
to
r
o
f
th
e
p
lan
e
it
r
esid
es
o
n
,
a
v
er
tical
s
am
p
lin
g
ax
is
is
co
n
s
tr
u
cted
alo
n
g
th
e
d
ir
e
ctio
n
o
f
th
e
n
o
r
m
al
v
ec
to
r
.
L
et
th
e
v
er
tical
s
am
p
lin
g
r
a
n
g
e
b
e
d
en
o
te
d
as
[
−
ℎ
]
an
d
th
e
s
am
p
lin
g
in
ter
v
al
as
ℎ
.
T
h
e
to
tal
n
u
m
b
er
o
f
s
am
p
lin
g
lay
er
s
is
g
iv
en
b
y
:
=
2
ℎ
ℎ
(
2
)
Fo
r
th
e
-
th
s
am
p
lin
g
lay
er
(
=
1
,
2
,
.
.
.
,
)
,
its
s
p
atial
p
o
s
itio
n
is
d
ef
in
ed
as
(
3
)
.
=
+
(
−
+
1
2
)
⋅
ℎ
⋅
(
3
)
T
h
is
f
o
r
m
u
la
e
n
s
u
r
es th
at
th
e
s
am
p
lin
g
lay
er
s
ar
e
s
y
m
m
et
r
ically
d
is
tr
ib
u
ted
ar
o
u
n
d
th
e
k
e
y
p
o
in
t
i
v
,
with
th
e
ce
n
tr
al
lay
er
(
=
+
1
2
)
co
r
r
esp
o
n
d
in
g
to
th
e
p
o
s
itio
n
o
f
th
e
k
ey
p
o
in
t,
t
h
u
s
p
r
eser
v
in
g
th
e
s
p
atial
s
y
m
m
etr
y
o
f
th
e
lo
ca
l in
te
n
s
ity
f
ea
tu
r
es.
2
.
3
.
2
.
I
nte
ns
it
y
f
ea
t
ure
a
g
g
re
g
a
t
io
n m
ec
ha
nis
m
At
ea
ch
s
am
p
lin
g
lay
er
,
a
s
p
h
er
ical
n
eig
h
b
o
r
h
o
o
d
with
r
ad
i
u
s
is
d
ef
in
ed
as:
=
{
∈
|
‖
−
‖
≤
}
(
4
)
wh
er
e
d
en
o
tes
th
e
in
p
u
t
p
o
i
n
t
clo
u
d
.
T
h
is
n
ei
g
h
b
o
r
h
o
o
d
d
esig
n
s
tr
ik
es
a
b
ala
n
ce
b
etwe
en
s
p
atial
r
eso
lu
tio
n
an
d
co
m
p
u
tatio
n
al
ef
f
icien
c
y
,
en
s
u
r
in
g
s
u
f
f
icien
t
lo
ca
l
in
t
en
s
ity
in
f
o
r
m
atio
n
is
ca
p
tu
r
e
d
wh
ile
m
itig
atin
g
n
o
is
e
in
tr
o
d
u
ce
d
b
y
o
v
er
ly
lar
g
e
n
eig
h
b
o
r
h
o
o
d
s
.
Fo
r
ea
ch
s
am
p
lin
g
la
y
er
,
th
e
m
ea
n
in
ten
s
ity
is
co
m
p
u
te
d
as
th
e
r
ep
r
esen
tativ
e
f
ea
t
u
r
e
(
5
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
52
-
62
56
=
{
1
|
|
∑
∈
(
)
,
if
|
|
>
0
0
,
if
|
|
=
0
(
5
)
(
)
r
ep
r
esen
ts
th
e
in
ten
s
ity
v
alu
e
o
f
p
o
in
t
.
A
ze
r
o
v
alu
e
is
ass
ig
n
ed
in
th
e
ca
s
e
o
f
an
e
m
p
ty
s
et,
en
s
u
r
in
g
d
escr
ip
to
r
co
m
p
leten
ess
an
d
p
r
ev
en
tin
g
n
u
m
e
r
ical
an
o
m
alie
s
d
u
r
in
g
c
o
m
p
u
tatio
n
.
2
.
3
.
3
.
L
o
ca
l
inte
ns
it
y
des
cr
ip
t
o
r
co
ns
t
ruct
io
n
Fin
ally
,
th
e
lo
ca
l in
ten
s
ity
d
es
cr
ip
to
r
f
o
r
a
k
ey
p
o
in
t
is
co
n
s
tr
u
cted
as
(
6
)
.
(
)
=
[
1
,
2
,
.
.
.
,
]
(
6
)
T
h
is
-
d
im
en
s
io
n
al
v
ec
to
r
d
escr
ip
to
r
e
f
f
ec
tiv
ely
ca
p
tu
r
es
th
e
v
er
tical
i
n
ten
s
ity
v
a
r
iatio
n
p
a
tter
n
ar
o
u
n
d
th
e
k
ey
p
o
in
t.
C
o
m
p
ar
e
d
to
tr
ad
it
io
n
al
s
in
g
le
-
p
o
in
t
i
n
ten
s
ity
v
a
lu
es,
th
is
lay
er
e
d
r
e
p
r
esen
tatio
n
p
r
o
v
id
es
a
m
o
r
e
co
m
p
r
eh
e
n
s
iv
e
ch
ar
ac
ter
izatio
n
o
f
th
e
lo
ca
l
m
ater
ial
p
r
o
p
er
ties
an
d
g
eo
m
etr
ic
s
tr
u
ctu
r
e,
o
f
f
er
in
g
r
ich
er
an
d
m
o
r
e
r
o
b
u
s
t in
ten
s
ity
in
f
o
r
m
at
io
n
f
o
r
s
u
b
s
eq
u
en
t
f
ea
tu
r
e
m
atch
in
g
.
2
.
4
.
I
ST
D
des
cr
ipto
r
def
ini
t
i
o
n
B
y
co
m
b
in
in
g
g
eo
m
et
r
ic
f
ea
tu
r
es a
n
d
in
ten
s
ity
f
ea
tu
r
es,
th
e
co
m
p
lete
I
STD
is
d
ef
in
ed
as
(
7
)
.
=
{
,
(
1
)
,
(
2
)
,
(
3
)
}
(
7
)
=
{
1
,
2
,
3
,
,
,
}
d
en
o
tes
th
e
o
r
ig
in
al
g
eo
m
etr
i
c
f
ea
tu
r
es
o
f
th
e
STD
d
escr
ip
t
o
r
an
d
(
)
ar
e
th
e
lo
ca
l
in
ten
s
ity
f
ea
tu
r
es
c
o
r
r
esp
o
n
d
i
n
g
to
th
e
th
r
ee
v
e
r
tices
o
f
th
e
tr
ian
g
le,
r
esp
ec
tiv
el
y
.
T
h
i
s
h
y
b
r
i
d
d
escr
ip
to
r
p
r
eser
v
es
th
e
s
p
atial
ex
p
r
ess
iv
en
ess
o
f
th
e
g
eo
m
etr
ic
f
ea
t
u
r
es
wh
ile
s
ig
n
if
ica
n
tly
e
n
h
a
n
cin
g
s
en
s
itiv
ity
t
o
m
ater
ial
p
r
o
p
e
r
ties
,
th
er
eb
y
im
p
r
o
v
i
n
g
f
ea
t
u
r
e
d
is
cr
im
in
ab
i
lity
in
g
eo
m
etr
ically
s
im
ilar
e
n
v
ir
o
n
m
en
ts
.
2
.
5
.
I
nte
ns
it
y
-
co
ns
is
t
ency
-
ba
s
ed
m
a
t
ching
v
er
if
ica
t
i
o
n
2
.
5
.
1
.
Sea
rc
h
lo
o
p c
a
nd
ida
t
e
Sin
ce
h
u
n
d
r
ed
s
o
f
d
escr
ip
to
r
s
ca
n
b
e
e
x
tr
ac
ted
f
r
o
m
a
s
in
g
le
k
ey
f
r
am
e,
we
em
p
l
o
y
a
h
a
s
h
tab
le
to
s
to
r
e
an
d
r
etr
iev
e
d
escr
ip
to
r
s
d
u
r
in
g
m
atch
i
n
g
ef
f
icien
tly
.
S
ix
attr
ib
u
tes
in
v
ar
ian
t
to
r
o
tatio
n
an
d
tr
an
s
latio
n
ar
e
u
s
ed
to
co
m
p
u
te
th
e
h
ash
k
ey
:
th
e
s
id
e
len
g
th
s
12
,
23
，
13
,
an
d
th
e
d
o
t
p
r
o
d
u
cts
o
f
t
h
e
p
r
o
jecte
d
n
o
r
m
al
v
ec
to
r
s
1
⋅
2
，
2
⋅
3
，
1
⋅
3
.
Descr
ip
to
r
s
s
h
ar
in
g
all
s
ix
s
im
ilar
attr
ib
u
tes
ar
e
ass
ig
n
ed
th
e
s
am
e
h
ash
k
ey
an
d
s
to
r
e
d
in
th
e
s
am
e
co
n
tain
er
.
Fo
r
a
q
u
e
r
y
k
ey
f
r
am
e,
all
d
escr
ip
to
r
s
ar
e
e
x
tr
ac
ted
.
Fo
r
ea
ch
d
escr
ip
to
r
,
its
h
ash
k
ey
is
co
m
p
u
ted
,
an
d
th
e
co
r
r
esp
o
n
d
in
g
co
n
tain
er
i
n
th
e
h
ash
tab
le
is
lo
ca
ted
.
E
ac
h
k
ey
f
r
am
e
th
at
h
as
at
least
o
n
e
d
escr
ip
to
r
in
th
e
m
atch
ed
c
o
n
tain
er
r
ec
eiv
es
o
n
e
v
o
te.
Af
t
er
all
d
escr
ip
to
r
s
f
r
o
m
th
e
q
u
er
y
k
e
y
f
r
am
e
ar
e
p
r
o
ce
s
s
ed
,
th
e
to
p
1
0
k
ey
f
r
a
m
es
with
th
e
h
ig
h
est
v
o
te
co
u
n
ts
ar
e
s
elec
ted
as
lo
o
p
clo
s
u
r
e
ca
n
d
id
ates.
T
h
e
m
atch
ed
d
escr
ip
to
r
s
ar
e
r
etain
ed
f
o
r
s
u
b
s
eq
u
en
t lo
o
p
d
etec
tio
n
.
2
.
5
.
2
.
I
nte
ns
it
y
f
ea
t
ure
s
im
ila
rit
y
m
ea
s
urem
ent
Du
r
in
g
th
e
t
r
ian
g
u
la
r
m
atch
in
g
p
r
o
ce
s
s
,
tr
ad
itio
n
al
g
eo
m
etr
ic
co
n
s
tr
ain
ts
ar
e
ef
f
ec
tiv
e
i
n
elim
in
atin
g
th
e
m
o
s
t
o
b
v
io
u
s
m
is
m
atch
es.
Ho
wev
er
,
th
ey
ar
e
lim
ited
in
s
ce
n
es
wh
er
e
th
e
g
eo
m
etr
ic
s
t
r
u
ctu
r
es
ar
e
s
im
ilar
b
u
t
d
if
f
e
r
in
m
ater
ial
p
r
o
p
e
r
ties
.
T
h
e
in
tr
o
d
u
ctio
n
o
f
in
ten
s
i
ty
f
ea
tu
r
es
p
r
o
v
id
es
a
r
eliab
le
b
asis
f
o
r
r
eso
lv
in
g
s
u
ch
am
b
ig
u
o
u
s
m
atch
es.
Fo
r
a
ca
n
d
id
ate
p
air
o
f
m
atch
e
d
tr
ian
g
les
(
,
)
,
th
e
co
n
s
is
ten
cy
o
f
in
ten
s
ity
f
ea
tu
r
es
b
etwe
en
c
o
r
r
esp
o
n
d
in
g
v
er
tices
is
ev
alu
ate
d
u
s
in
g
co
s
in
e
s
im
ilar
ity
.
Fo
r
th
e
i
-
th
p
air
o
f
co
r
r
esp
o
n
d
in
g
v
er
tices,
th
e
in
t
en
s
ity
f
ea
tu
r
e
s
im
ilar
ity
is
d
ef
i
n
ed
as
(
8
)
.
(
)
=
(
)
⋅
(
)
‖
(
)
‖
⋅
‖
(
)
‖
(
8
)
T
h
e
ch
o
ice
o
f
co
s
in
e
s
im
ilar
ity
o
f
f
er
s
t
h
e
f
o
llo
win
g
th
e
o
r
etica
l
ad
v
an
tag
es:
First,
it
i
s
in
v
ar
ian
t
to
th
e
ab
s
o
lu
te
s
ca
le
o
f
in
ten
s
ity
v
al
u
es,
m
ak
in
g
it
r
o
b
u
s
t
to
ca
lib
r
atio
n
d
if
f
er
e
n
ce
s
ac
r
o
s
s
d
if
f
er
en
t
L
iDAR
s
en
s
o
r
s
.
Seco
n
d
,
co
s
in
e
s
im
ilar
ity
em
p
h
asizes
th
e
p
atter
n
o
f
in
ten
s
ity
v
ar
iatio
n
r
ath
e
r
th
an
ab
s
o
lu
te
v
alu
es,
wh
ich
alig
n
s
with
o
u
r
d
esig
n
g
o
al
o
f
ca
p
tu
r
i
n
g
lo
ca
l
m
ate
r
ial
ch
ar
ac
ter
is
tics
.
Fin
ally
,
it
is
co
m
p
u
tatio
n
ally
ef
f
icien
t
,
m
ak
in
g
it
s
u
itab
le
f
o
r
r
ea
l
-
ti
m
e
ap
p
licatio
n
s
.
T
o
ev
alu
ate
th
e
o
v
er
all
in
ten
s
ity
co
n
s
is
ten
cy
o
f
a
tr
ian
g
le,
th
e
ag
g
r
eg
ated
in
ten
s
ity
s
im
ilar
ity
is
d
ef
in
ed
as
(
9
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
I
S
TD
-
LIO
M
:
Dir
ec
t
LiDAR
-
in
erti
a
l o
d
o
metry
a
n
d
ma
p
p
in
g
w
ith
in
ten
s
ity
-
en
h
a
n
ce
d
s
ta
b
le
…
(
Lixia
o
Ya
n
g
)
57
=
1
3
∑
(
)
3
=
1
(
9
)
2
.
5
.
3
.
I
nte
ns
it
y
v
er
if
ica
t
io
n c
rit
er
io
n
T
h
e
in
ten
s
ity
-
b
ased
m
atch
in
g
v
er
if
icatio
n
em
p
lo
y
s
a
d
u
al
-
th
r
esh
o
ld
s
tr
ateg
y
,
en
s
u
r
in
g
b
o
th
o
v
er
al
l
co
n
s
is
ten
cy
an
d
lo
ca
l
r
eliab
i
lity
.
T
h
e
g
lo
b
al
co
n
s
is
ten
cy
co
n
s
tr
ain
t
r
eq
u
ir
es
th
at
th
e
av
er
ag
e
i
n
ten
s
ity
s
im
ilar
ity
o
f
th
e
tr
ian
g
le
s
atis
f
ies:
>
(
1
0
)
T
h
e
lo
ca
l r
eliab
ilit
y
co
n
s
tr
ain
t r
eq
u
ir
es th
at
th
e
i
n
ten
s
ity
s
im
ilar
ity
o
f
ea
ch
v
er
tex
s
atis
f
ies:
=
1
,
2
,
3
(
)
>
(
1
1
)
wh
er
e
is
th
e
g
lo
b
al
in
ten
s
ity
s
im
ilar
ity
th
r
esh
o
ld
an
d
is
t
h
e
m
in
im
u
m
s
im
ilar
ity
th
r
esh
o
ld
f
o
r
ea
ch
in
d
iv
id
u
al
v
er
tex
.
2
.
6
.
L
o
o
p
clo
s
ure
det
ec
t
io
n
a
nd
g
lo
ba
l po
s
e
g
ra
ph
o
ptimiza
t
io
n
2
.
6
.
1
.
L
o
o
p
clo
s
ure
det
ec
t
io
n
T
h
is
p
ap
er
a
d
o
p
ts
a
lo
o
p
c
lo
s
u
r
e
d
etec
tio
n
a
p
p
r
o
ac
h
th
at
co
m
b
in
es
a
d
is
tan
ce
-
th
r
esh
o
ld
-
b
ase
d
k
ey
f
r
am
e
s
elec
tio
n
s
tr
ateg
y
w
ith
a
d
u
al
-
p
ath
ca
n
d
i
d
ate
s
cr
e
en
in
g
m
ec
h
an
is
m
.
T
h
e
s
y
s
te
m
s
elec
ts
k
ey
f
r
am
es
b
ased
o
n
a
p
r
ed
e
f
in
ed
s
p
ati
al
d
is
tan
ce
th
r
esh
o
ld
ℎ
.
W
h
en
th
e
r
o
b
o
t'
s
tr
av
eled
d
is
tan
ce
ex
ce
ed
s
th
is
th
r
esh
o
ld
,
th
e
cu
r
r
en
t
f
r
a
m
e
is
d
esig
n
ated
as
a
n
ew
k
ey
f
r
am
e.
All
f
r
am
es
with
in
th
is
th
r
es
h
o
ld
ar
e
ass
o
ciate
d
with
th
eir
co
r
r
esp
o
n
d
in
g
k
e
y
f
r
am
e,
f
o
r
m
in
g
a
p
o
in
t
clo
u
d
s
et
f
o
r
ea
ch
k
ey
f
r
a
m
e.
Du
r
i
n
g
lo
o
p
ca
n
d
id
ate
s
elec
tio
n
,
two
c
o
m
p
lem
en
ta
r
y
f
ilter
in
g
p
ath
s
ar
e
e
m
p
lo
y
e
d
:
i
)
f
ea
tu
r
e
-
b
ased
s
cr
ee
n
i
n
g
,
w
h
ich
s
elec
ts
th
e
t
o
p
h
is
to
r
ical
k
e
y
f
r
am
es
with
th
e
h
ig
h
est
s
im
ilar
ity
t
o
th
e
cu
r
r
e
n
t
k
ey
f
r
am
e
b
ased
o
n
th
e
I
STD
d
escr
ip
to
r
;
an
d
ii)
s
p
atial
p
r
o
x
im
ity
-
b
ased
s
cr
ee
n
in
g
,
wh
ic
h
s
elec
ts
th
e
to
p
h
is
to
r
ical
k
ey
f
r
a
m
es
th
at
ar
e
clo
s
est
in
E
u
clid
ea
n
s
p
ac
e
to
t
h
e
cu
r
r
en
t
k
ey
f
r
a
m
e'
s
p
o
s
e.
L
et
th
e
cu
r
r
en
t
k
e
y
f
r
am
e
p
o
s
e
b
e
,
an
d
th
e
s
et
o
f
h
is
to
r
ical
k
ey
f
r
am
e
p
o
s
es
b
e
{
1
,
2
,
3
,
.
.
.
,
−
1
}
.
T
h
en
,
th
e
s
p
atial
p
r
o
x
im
ity
-
b
a
s
ed
ca
n
d
id
ate
s
et
is
d
ef
in
ed
as
:
2
=
<
1
≪
‖
−
‖
2
,
|
2
|
=
(
1
2
)
wh
er
e
an
d
d
en
o
te
th
e
p
o
s
itio
n
v
ec
to
r
s
o
f
k
ey
f
r
am
es
an
d
,
r
esp
ec
tiv
ely
.
Fo
r
t
h
e
s
elec
ted
ca
n
d
id
ate
k
ey
f
r
am
es,
th
e
s
y
s
tem
p
er
f
o
r
m
s
p
r
ec
is
e
r
eg
is
tr
atio
n
u
s
in
g
a
p
o
in
t
-
to
-
p
lan
e
I
C
P
alg
o
r
ith
m
,
wh
ich
m
in
im
izes
th
e
p
o
in
t
-
to
-
p
lan
e
d
is
tan
ce
s
iter
ativ
ely
to
esti
m
ate
th
e
o
p
tim
al
r
elativ
e
tr
an
s
f
o
r
m
atio
n
.
I
n
th
e
g
eo
m
etr
ic
v
er
if
icatio
n
s
tag
e,
th
e
s
y
s
tem
co
m
p
r
e
h
en
s
iv
ely
e
v
alu
ates
i
n
d
icato
r
s
s
u
ch
as
r
eg
is
tr
atio
n
er
r
o
r
,
p
o
in
t
clo
u
d
o
v
er
lap
r
atio
,
an
d
tr
a
n
s
f
o
r
m
a
tio
n
p
lau
s
ib
ilit
y
to
id
en
tify
c
an
d
id
ates
with
h
ig
h
g
e
o
m
etr
i
c
co
n
s
is
ten
cy
.
T
h
e
f
in
al
lo
o
p
co
n
s
tr
ain
t
is
estab
li
s
h
ed
u
s
in
g
th
e
ca
n
d
i
d
ate
with
th
e
m
in
im
u
m
r
e
g
is
tr
atio
n
er
r
o
r
th
at
also
m
ee
ts
p
r
ed
ef
in
e
d
q
u
ality
th
r
esh
o
ld
s
,
an
d
its
r
elativ
e
p
o
s
e
tr
an
s
f
o
r
m
atio
n
is
in
co
r
p
o
r
ated
in
to
b
ac
k
-
en
d
o
p
tim
izatio
n
.
2
.
6
.
2
.
F
a
ct
o
r
g
r
a
ph
o
ptim
iza
t
io
n
I
n
th
is
p
ap
er
,
a
f
ac
to
r
g
r
ap
h
f
r
am
ewo
r
k
is
em
p
lo
y
e
d
f
o
r
g
lo
b
al
p
o
s
e
o
p
tim
izatio
n
,
f
o
r
m
u
latin
g
a
n
o
n
lin
ea
r
o
p
tim
izatio
n
p
r
o
b
le
m
th
at
in
co
r
p
o
r
ates
o
d
o
m
etr
y
f
ac
to
r
s
an
d
lo
o
p
clo
s
u
r
e
f
ac
to
r
s
.
I
n
th
e
f
ac
to
r
g
r
ap
h
r
ep
r
esen
tatio
n
,
r
elativ
e
p
o
s
e
co
n
s
tr
ain
ts
b
etwe
en
ad
ja
ce
n
t
k
ey
f
r
am
es
ar
e
en
c
o
d
ed
a
s
o
d
o
m
etr
y
f
ac
to
r
s
,
wh
ile
co
n
s
tr
ain
ts
o
b
tain
ed
f
r
o
m
lo
o
p
clo
s
u
r
e
d
etec
tio
n
ar
e
en
co
d
ed
as
lo
o
p
clo
s
u
r
e
f
a
cto
r
s
.
L
et
th
e
s
et
o
f
k
ey
f
r
am
e
p
o
s
es
b
e
d
en
o
ted
as
=
{
1
,
2
,
.
.
.
,
}
,
th
e
s
et
o
f
o
d
o
m
etr
y
co
n
s
tr
ain
ts
as
=
{
,
+
1
}
,
an
d
th
e
s
et
o
f
lo
o
p
cl
o
s
u
r
e
co
n
s
tr
ain
ts
as
ℒ
=
{
,
}
.
T
h
e
f
ac
to
r
g
r
a
p
h
o
p
tim
iza
tio
n
p
r
o
b
lem
ca
n
th
en
b
e
f
o
r
m
u
lated
as
(
1
3
)
.
∗
=
a
r
g
min
[
∑
−
1
=
1
‖
−
1
⋅
+
1
−
,
+
1
‖
∑
2
+
∑
(
,
)
∈
ℒ
‖
−
1
⋅
−
,
‖
∑
2
]
(
1
3
)
H
er
e,
th
e
f
ir
s
t
ter
m
r
ep
r
esen
t
s
th
e
o
d
o
m
etr
y
f
ac
to
r
s
,
wh
ic
h
co
n
s
tr
ain
th
e
r
elativ
e
p
o
s
es
b
etwe
en
a
d
jace
n
t
k
ey
f
r
am
es,
an
d
th
e
s
ec
o
n
d
t
er
m
r
e
p
r
esen
ts
th
e
lo
o
p
clo
s
u
r
e
f
ac
t
o
r
s
,
wh
ic
h
co
n
s
tr
ain
th
e
r
elativ
e
p
o
s
es
b
etwe
en
lo
o
p
-
clo
s
u
r
e
k
e
y
f
r
a
m
e
p
air
s
.
∑
an
d
d
en
o
te
th
e
c
o
v
ar
ian
ce
m
at
r
ices
ass
o
ciate
d
with
o
d
o
m
etr
y
an
d
lo
o
p
clo
s
u
r
e
o
b
s
er
v
atio
n
s
,
r
esp
ec
tiv
ely
.
T
h
e
s
y
s
tem
s
o
lv
es
th
e
o
p
tim
izatio
n
u
s
i
n
g
th
e
L
e
v
en
b
er
g
–
Ma
r
q
u
ar
d
t
alg
o
r
ith
m
,
wh
ich
a
d
ap
tiv
ely
ad
ju
s
ts
th
e
d
am
p
in
g
f
ac
to
r
to
s
tr
ik
e
a
b
alan
ce
b
etwe
en
co
n
v
er
g
e
n
ce
s
p
ee
d
an
d
s
tab
ilit
y
.
An
in
c
r
em
en
tal
u
p
d
ate
m
ec
h
an
is
m
i
s
also
im
p
lem
en
ted
:
wh
en
a
n
ew
lo
o
p
cl
o
s
u
r
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
52
-
62
58
co
n
s
tr
ain
t
is
d
etec
ted
,
o
p
tim
izatio
n
is
p
er
f
o
r
m
e
d
o
n
ly
o
n
th
e
lo
ca
l
s
u
b
g
r
ap
h
,
s
ig
n
if
ican
tly
im
p
r
o
v
in
g
co
m
p
u
tatio
n
al
ef
f
icien
c
y
.
Af
ter
o
p
tim
izatio
n
,
t
h
e
s
y
s
tem
r
e
-
r
eg
is
ter
s
th
e
p
o
in
t
clo
u
d
s
b
as
ed
o
n
th
e
co
r
r
ec
ted
p
o
s
es to
g
en
er
ate
a
g
lo
b
ally
c
o
n
s
is
ten
t m
ap
o
f
th
e
en
v
ir
o
n
m
en
t.
2
.
6
.
3
.
M
a
p
c
o
ns
t
ruct
io
n
T
h
e
m
ap
co
n
s
tr
u
ctio
n
m
o
d
u
l
e
ap
p
lies
th
e
g
l
o
b
ally
c
o
n
s
is
ten
t
p
o
s
es
o
b
tain
ed
f
r
o
m
f
ac
to
r
g
r
a
p
h
o
p
tim
izatio
n
t
o
th
e
co
n
ca
ten
at
io
n
o
f
k
ey
f
r
a
m
e
p
o
in
t
clo
u
d
s
,
r
esu
ltin
g
in
a
h
ig
h
-
p
r
ec
is
io
n
3
D
r
ep
r
esen
tatio
n
o
f
th
e
en
v
ir
o
n
m
e
n
t.
L
et
th
e
lo
ca
l
p
o
in
t
clo
u
d
o
f
th
e
-
th
k
ey
f
r
am
e
b
e
d
en
o
ted
as
,
an
d
its
o
p
tim
ized
g
lo
b
al
p
o
s
e
as
∗
.
T
h
en
,
th
e
co
n
s
tr
u
ctio
n
o
f
th
e
g
lo
b
al
m
ap
ca
n
b
e
e
x
p
r
ess
ed
as
(
1
4
)
.
=
⋃
∗
⋅
=
1
(
1
4
)
H
er
e,
∗
⋅
r
ep
r
esen
ts
th
e
tr
a
n
s
f
o
r
m
atio
n
o
f
th
e
lo
ca
l
p
o
in
t
clo
u
d
in
to
th
e
g
lo
b
al
co
o
r
d
in
ate
s
y
s
tem
.
T
h
is
p
r
o
ce
s
s
alig
n
s
o
r
ig
in
ally
d
r
i
f
ted
lo
ca
l
p
o
i
n
t
clo
u
d
s
eg
m
en
ts
to
a
u
n
if
ied
g
lo
b
al
f
r
am
e,
ef
f
ec
tiv
ely
elim
in
atin
g
th
e
n
eg
ativ
e
im
p
ac
t
o
f
ac
c
u
m
u
lated
p
o
s
e
er
r
o
r
s
o
n
m
ap
co
n
s
is
ten
cy
.
T
o
en
s
u
r
e
m
ap
q
u
ality
,
th
e
s
y
s
tem
em
p
lo
y
s
ad
ap
tiv
e
v
o
x
el
f
ilter
i
n
g
f
o
r
d
o
w
n
s
am
p
lin
g
th
e
m
er
g
ed
p
o
i
n
t c
lo
u
d
:
=
(
,
)
(
1
5
)
wh
er
e
d
en
o
tes
t
h
e
ad
a
p
tiv
el
y
ad
ju
s
ted
v
o
x
el
s
ize,
wh
ich
r
ed
u
ce
s
r
ed
u
n
d
an
t
d
ata
w
h
ile
p
r
eser
v
in
g
ess
en
tial g
eo
m
etr
ic
f
ea
tu
r
es.
3.
RE
SU
L
T
S
AND
D
I
SCU
SS
I
O
N
3
.
1
.
P
re
cisi
o
n
-
re
ca
ll e
v
a
lua
t
io
n
I
n
th
is
p
ap
er
,
t
h
e
p
r
ec
is
io
n
-
r
e
ca
ll
cu
r
v
es
o
f
th
r
ee
d
escr
ip
to
r
s
ar
e
ev
alu
ated
u
s
in
g
s
eq
u
en
c
es
0
5
,
0
6
,
an
d
0
8
f
r
o
m
th
e
KI
T
T
I
d
ataset.
Gr
o
u
n
d
-
tr
u
th
lo
o
p
clo
s
u
r
es
ar
e
m
an
u
ally
an
n
o
tated
f
o
r
th
ese
s
eq
u
en
ce
s
.
T
h
e
an
n
o
tatio
n
r
esu
lts
ar
e
s
h
o
wn
in
Fig
u
r
e
3
,
wh
er
e
b
lu
e
p
o
s
es
r
ep
r
esen
t
lo
o
p
clo
s
u
r
e
f
r
am
es
th
at
r
ev
is
it
p
r
ev
io
u
s
ly
e
x
p
lo
r
e
d
lo
ca
tio
n
s
.
Fig
u
r
e
3
.
Vis
u
aliza
tio
n
o
f
KI
T
T
I
s
eq
u
en
ce
s
Af
ter
g
r
o
u
n
d
-
tr
u
th
an
n
o
tatio
n
,
th
e
p
r
o
p
o
s
ed
I
STD
d
escr
ip
to
r
is
ev
alu
ated
o
n
th
e
th
r
ee
d
atasets
b
y
p
lo
ttin
g
its
p
r
ec
is
io
n
-
r
ec
all
cu
r
v
e.
Mu
ltip
le
p
r
ec
is
io
n
a
n
d
r
e
ca
ll
v
alu
es
ar
e
o
b
tain
ed
b
y
ad
ju
s
tin
g
p
ar
am
eter
s
to
co
n
s
tr
u
ct
th
e
p
r
ec
is
io
n
-
r
ec
a
ll
cu
r
v
e.
Fo
r
ea
ch
s
eq
u
en
ce
,
if
a
d
etec
ted
lo
o
p
clo
s
u
r
e
lies
with
in
2
0
m
eter
s
o
f
th
e
q
u
er
y
f
r
am
e,
it
is
m
ar
k
e
d
as
a
tr
u
e
p
o
s
itiv
e
(
T
P)
;
if
n
o
lo
o
p
cl
o
s
u
r
e
is
d
etec
ted
a
t
a
n
an
n
o
tated
lo
o
p
clo
s
u
r
e
f
r
am
e,
it
is
m
ar
k
ed
as
a
f
alse
n
eg
ativ
e
(
FN)
;
an
d
if
a
d
etec
ted
lo
o
p
clo
s
u
r
e
d
o
es
n
o
t
co
r
r
esp
o
n
d
th
e
q
u
er
y
f
r
am
e
with
in
2
0
m
eter
s
,
it
is
m
ar
k
ed
as
a
f
alse
p
o
s
itiv
e
(
FP
)
.
p
r
ec
is
io
n
(
P
)
an
d
r
ec
all
(
R
)
ar
e
th
e
n
ca
lcu
lated
as
(
1
6
)
.
=
+
,
=
+
(
1
6
)
T
o
v
alid
ate
th
e
ef
f
ec
tiv
e
n
ess
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
,
it
i
s
co
m
p
ar
ed
with
s
ca
n
co
n
tex
t
[1
2
]
an
d
STD
[
15
]
o
n
th
r
ee
s
eq
u
en
ce
s
o
f
th
e
p
u
b
licly
av
ailab
le
KI
T
T
I
d
ataset.
T
h
e
ex
p
er
im
en
tal
r
esu
lts
o
n
th
e
th
r
ee
s
eq
u
en
ce
s
ar
e
s
h
o
wn
in
Fig
u
r
e
4
.
I
t
ca
n
b
e
o
b
s
er
v
e
d
th
at
th
e
p
r
o
p
o
s
ed
I
STD
d
escr
ip
to
r
ac
h
iev
es
a
h
i
g
h
er
ar
ea
u
n
d
er
t
h
e
cu
r
v
e
(
AUC),
in
d
ica
tin
g
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
I
S
TD
-
LIO
M
:
Dir
ec
t
LiDAR
-
in
erti
a
l o
d
o
metry
a
n
d
ma
p
p
in
g
w
ith
in
ten
s
ity
-
en
h
a
n
ce
d
s
ta
b
le
…
(
Lixia
o
Ya
n
g
)
59
Fig
u
r
e
4
.
Per
f
o
r
m
an
c
e
c
o
m
p
ar
is
o
n
o
n
th
e
KI
T
T
I
d
ataset
3
.
2
.
Q
ua
ntit
a
t
iv
e
c
o
m
pa
riso
n o
f
lo
ca
liza
t
i
o
n a
cc
ura
cy
T
o
f
u
r
th
er
v
er
if
y
th
e
ac
cu
r
ac
y
im
p
r
o
v
em
en
t
b
r
o
u
g
h
t
b
y
I
STD
to
g
lo
b
al
o
p
tim
izatio
n
with
in
th
e
o
v
er
all
s
y
s
tem
,
a
q
u
an
titativ
e
ev
alu
atio
n
o
f
l
o
ca
lizatio
n
ac
cu
r
ac
y
is
co
n
d
u
cted
.
As
s
h
o
wn
in
Fig
u
r
e
5
,
th
e
r
esu
lts
ar
e
o
b
tain
ed
o
n
th
r
e
e
test
s
eq
u
en
ce
s
f
r
o
m
th
e
KI
T
T
I
d
ataset
(
Seq
u
en
ce
s
0
5
,
0
7
,
an
d
0
8
)
.
T
h
e
ev
alu
ated
tr
ajec
to
r
ies
in
clu
d
e
th
e
FAST
-
L
I
O2
[
1
9
]
p
o
s
es
L
I
O,
th
e
b
ac
k
-
en
d
o
p
tim
izati
o
n
b
ased
o
n
STD
(
l
oop
-
STD)
,
a
n
d
th
e
b
ac
k
-
en
d
o
p
tim
izatio
n
b
ased
o
n
I
S
T
D
(
L
o
o
p
-
I
STD)
.
T
h
e
r
e
d
b
o
x
es
h
ig
h
lig
h
t
th
e
d
ev
iatio
n
o
f
th
e
tr
ajec
to
r
y
en
d
p
o
in
ts
f
r
o
m
th
e
g
r
o
u
n
d
tr
u
th
f
o
r
ea
ch
m
eth
o
d
.
I
t
ca
n
b
e
o
b
s
e
r
v
ed
th
at
o
d
o
m
etr
y
ex
h
ib
its
s
ig
n
if
ican
t
d
r
if
t
in
la
r
g
e
-
s
ca
le
s
ce
n
ar
io
s
,
wh
ile
th
e
tr
ajec
to
r
y
o
p
tim
ized
u
s
in
g
I
S
T
D
is
clo
s
er
to
th
e
g
r
o
u
n
d
tr
u
th
c
o
m
p
ar
e
d
to
th
a
t
o
f
STD,
in
d
icatin
g
h
ig
h
er
l
o
ca
lizatio
n
ac
cu
r
ac
y
.
T
h
e
q
u
a
n
titativ
e
r
esu
lts
ar
e
p
r
esen
ted
in
T
a
b
le
1
,
wh
e
r
e
I
S
T
D
ac
h
iev
es th
e
lo
west tr
ajec
to
r
y
er
r
o
r
ac
r
o
s
s
all
th
r
ee
test
ca
s
es.
Fig
u
r
e
5
.
Vis
u
al
c
o
m
p
ar
is
o
n
o
f
KI
T
T
I
s
eq
u
e
n
ce
s
T
ab
le
1
.
Ab
s
o
lu
te
p
o
s
itio
n
er
r
o
r
s
in
KI
T
T
I
Data
s
ets with
d
if
f
er
en
t m
eth
o
d
s
M
e
t
h
o
d
R
mse
mea
n
m
e
d
i
a
n
K
I
TTI
0
5
K
I
TTI
0
7
K
I
TTI
0
8
F
A
S
T
-
LI
O
2
2
.
1
6
1
.
9
9
1
.
8
9
0
.
8
4
0
.
7
1
0
.
6
5
4
.
8
3
4
.
1
4
3
.
9
4
Lo
o
p
-
S
TD
1
.
4
6
1
.
3
2
1
.
3
3
0
.
4
5
0
.
3
9
0
.
4
3
7
.
0
3
6
.
1
5
6
.
2
0
L
o
o
p
-
i
S
T
D
1
.
2
8
1
.
1
6
1
.
1
7
0
.
4
3
0
.
3
8
0
.
4
2
4
.
2
0
3
.
6
0
3
.
3
4
3
.
3
.
Q
ua
lit
a
t
iv
e
c
o
m
pa
riso
n
o
f
ma
pp
ing
re
s
ults
As
illu
s
tr
ated
in
Fig
u
r
e
6
,
th
e
m
ap
p
in
g
r
esu
lts
f
o
r
KI
T
T
I
s
eq
u
en
ce
s
0
5
,
0
7
,
an
d
0
8
ar
e
p
r
esen
ted
to
q
u
alitativ
ely
co
m
p
ar
e
th
e
e
f
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
.
I
n
ea
ch
s
u
b
f
i
g
u
r
e,
t
h
e
r
e
d
d
ash
e
d
b
o
x
lo
ca
ted
in
th
e
l
o
wer
lef
t
co
r
n
e
r
h
ig
h
lig
h
ts
th
e
p
o
i
n
t
clo
u
d
m
ap
co
n
s
tr
u
cted
u
s
in
g
r
aw
o
d
o
m
etr
y
p
o
s
es.
Du
e
t
o
th
e
ac
cu
m
u
latio
n
o
f
p
o
s
e
d
r
if
t
o
v
er
tim
e,
th
e
m
ap
s
g
e
n
er
ated
in
t
h
is
m
an
n
er
ex
h
i
b
it
ev
id
en
t
d
u
p
licate
s
h
ad
o
ws
an
d
m
is
alig
n
m
en
t,
wh
ich
n
eg
ativ
el
y
im
p
ac
t
o
v
e
r
all
m
ap
p
in
g
q
u
ality
.
I
n
co
n
t
r
ast,
th
e
lo
wer
r
ig
h
t
s
u
b
f
ig
u
r
e
s
h
o
ws
th
e
m
ap
p
i
n
g
r
esu
lt
o
b
tain
e
d
u
s
in
g
g
lo
b
ally
c
o
n
s
is
ten
t
p
o
s
es
r
ef
in
ed
b
y
b
ac
k
-
e
n
d
o
p
tim
izatio
n
.
T
h
e
u
s
e
o
f
o
p
ti
m
ized
p
o
s
es
s
ig
n
if
ican
tly
r
e
d
u
ce
s
d
u
p
licate
s
h
a
d
o
w
e
f
f
ec
ts
.
I
t
p
r
o
d
u
ce
s
a
m
o
r
e
co
h
er
en
t
an
d
ac
cu
r
ate
g
l
o
b
al
m
ap
,
d
em
o
n
s
tr
atin
g
th
e
ad
v
an
tag
e
o
f
th
e
p
r
o
p
o
s
ed
I
STD
-
b
a
s
ed
lo
o
p
clo
s
u
r
e
an
d
g
lo
b
al
o
p
tim
izatio
n
in
m
ain
tai
n
in
g
g
l
o
b
al
co
n
s
is
ten
cy
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
1
5
,
No
.
1
,
Ma
r
ch
20
2
6
:
52
-
62
60
Fig
u
r
e
6
.
C
o
m
p
a
r
is
o
n
o
f
m
ap
p
in
g
r
esu
lts
b
ef
o
r
e
an
d
af
ter
o
p
tim
izatio
n
4.
CO
NCLU
SI
O
N
T
h
is
s
tu
d
y
ad
d
r
ess
es
th
e
is
s
u
e
o
f
cu
m
u
lativ
e
d
r
if
t
in
lo
n
g
-
s
eq
u
e
n
ce
task
s
in
v
o
lv
in
g
LIO
b
y
p
r
o
p
o
s
in
g
a
lo
o
p
clo
s
u
r
e
d
ete
ctio
n
f
r
am
ewo
r
k
b
ased
o
n
in
t
en
s
ity
-
en
h
an
ce
d
d
escr
ip
to
r
,
c
alled
I
STD
-
L
I
OM
.
B
y
in
teg
r
atin
g
p
o
in
t
clo
u
d
i
n
ten
s
ity
d
is
tr
ib
u
tio
n
with
g
e
o
m
etr
ic
s
p
atial
f
ea
tu
r
es,
we
d
esig
n
ed
a
r
o
tatio
n
-
in
v
ar
ian
t
I
STD
d
escr
ip
to
r
.
C
o
m
b
in
ed
with
a
d
y
n
am
ic
k
e
y
f
r
am
e
tr
ig
g
er
i
n
g
m
ec
h
an
is
m
an
d
a
h
y
b
r
i
d
lo
o
p
clo
s
u
r
e
r
etr
iev
al
s
tr
ateg
y
,
th
i
s
ap
p
r
o
ac
h
e
n
ab
les
th
e
s
y
s
tem
atic
in
co
r
p
o
r
atio
n
o
f
in
te
n
s
ity
in
f
o
r
m
atio
n
in
to
lo
o
p
clo
s
u
r
e
d
etec
tio
n
.
E
x
p
e
r
im
en
tal
r
esu
lts
d
em
o
n
s
tr
ate
th
at
th
e
m
eth
o
d
s
ig
n
if
ican
tly
e
n
h
an
ce
s
lo
o
p
clo
s
u
r
e
r
o
b
u
s
tn
ess
an
d
g
lo
b
al
m
ap
co
n
s
is
ten
cy
o
n
th
e
KI
T
T
I
d
ataset,
v
alid
atin
g
th
e
ef
f
ec
tiv
e
n
ess
o
f
in
ten
s
ity
f
ea
t
u
r
es
in
co
r
r
ec
tin
g
L
iDAR
-
in
er
tial
o
d
o
m
etr
y
d
r
if
t.
H
o
wev
er
,
th
e
cu
r
r
en
t
r
esear
c
h
s
till
h
as
lim
itatio
n
s
in
d
y
n
am
ic
s
ce
n
e
ad
ap
tab
ilit
y
an
d
m
u
lt
i
-
s
en
s
o
r
co
llab
o
r
atio
n
.
Fu
tu
r
e
wo
r
k
will
f
o
c
u
s
o
n
d
esig
n
in
g
s
em
a
n
tically
en
h
an
ce
d
d
escr
ip
to
r
s
an
d
c
r
o
s
s
-
m
o
d
al
d
ata
f
u
s
io
n
t
o
f
u
r
t
h
er
en
h
an
ce
th
e
s
y
s
tem
'
s
g
en
er
aliza
tio
n
in
co
m
p
lex
en
v
ir
o
n
m
en
ts
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
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
L
ix
iao
Yan
g
✓
✓
✓
✓
✓
✓
✓
✓
✓
Sh
en
g
Hu
a
✓
✓
Yo
u
b
in
g
Fen
g
✓
✓
✓
✓
Sh
an
g
zo
n
g
Yan
g
✓
✓
J
ie
W
an
g
✓
✓
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.
DATA AV
AI
L
AB
I
L
I
T
Y
Der
iv
ed
d
ata
s
u
p
p
o
r
ti
n
g
th
e
f
in
d
in
g
s
o
f
th
is
s
tu
d
y
ar
e
av
ailab
le
f
r
o
m
th
e
co
r
r
esp
o
n
d
i
n
g
au
th
o
r
[
FY
]
o
n
r
e
q
u
est
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
I
S
TD
-
LIO
M
:
Dir
ec
t
LiDAR
-
in
erti
a
l o
d
o
metry
a
n
d
ma
p
p
in
g
w
ith
in
ten
s
ity
-
en
h
a
n
ce
d
s
ta
b
le
…
(
Lixia
o
Ya
n
g
)
61
RE
F
E
R
E
NC
E
S
[
1
]
W
.
X
u
a
n
d
F
.
Z
h
a
n
g
,
“
F
A
S
T
-
LI
O
:
a
f
a
s
t
,
r
o
b
u
st
L
i
D
A
R
-
i
n
e
r
t
i
a
l
o
d
o
me
t
r
y
p
a
c
k
a
g
e
b
y
t
i
g
h
t
l
y
-
c
o
u
p
l
e
d
i
t
e
r
a
t
e
d
K
a
l
ma
n
f
i
l
t
e
r
,
”
I
EEE
R
o
b
o
t
i
c
s
a
n
d
Au
t
o
m
a
t
i
o
n
L
e
t
t
e
r
s
,
v
o
l
.
6
,
n
o
.
2
,
p
p
.
3
3
1
7
–
3
3
2
4
,
A
p
r
.
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
L
R
A
.
2
0
2
1
.
3
0
6
4
2
2
7
.
[
2
]
Y
.
P
a
n
,
P
.
X
i
a
o
,
Y
.
H
e
,
Z.
S
h
a
o
,
a
n
d
Z.
L
i
,
“
M
U
LLS
:
v
e
r
s
a
t
i
l
e
L
i
D
A
R
S
L
A
M
v
i
a
m
u
l
t
i
-
met
r
i
c
l
i
n
e
a
r
l
e
a
st
s
q
u
a
r
e
,
”
i
n
2
0
2
1
I
EEE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
R
o
b
o
t
i
c
s
a
n
d
A
u
t
o
m
a
t
i
o
n
(
I
C
R
A)
,
I
EEE,
M
a
y
2
0
2
1
,
p
p
.
1
1
6
3
3
–
1
1
6
4
0
.
d
o
i
:
1
0
.
1
1
0
9
/
I
C
R
A
4
8
5
0
6
.
2
0
2
1
.
9
5
6
1
3
6
4
.
[
3
]
C
.
B
a
i
,
T
.
X
i
a
o
,
Y
.
C
h
e
n
,
H
.
W
a
n
g
,
F
.
Z
h
a
n
g
,
a
n
d
X
.
G
a
o
,
“
F
a
s
t
e
r
-
LI
O
:
l
i
g
h
t
w
e
i
g
h
t
t
i
g
h
t
l
y
c
o
u
p
l
e
d
Li
D
A
R
-
i
n
e
r
t
i
a
l
o
d
o
met
r
y
u
si
n
g
p
a
r
a
l
l
e
l
sp
a
r
se
i
n
c
r
e
m
e
n
t
a
l
v
o
x
e
l
s,
”
I
EE
E
R
o
b
o
t
i
c
s
a
n
d
A
u
t
o
m
a
t
i
o
n
L
e
t
t
e
rs
,
v
o
l
.
7
,
n
o
.
2
,
p
p
.
4
8
6
1
–
4
8
6
8
,
A
p
r
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
L
R
A
.
2
0
2
2
.
3
1
5
2
8
3
0
.
[
4
]
D
.
H
e
,
W
.
X
u
,
N
.
C
h
e
n
,
F
.
K
o
n
g
,
C
.
Y
u
a
n
,
a
n
d
F
.
Z
h
a
n
g
,
“
P
o
i
n
t
‐
LI
O
:
r
o
b
u
st
h
i
g
h
‐
b
a
n
d
w
i
d
t
h
l
i
g
h
t
d
e
t
e
c
t
i
o
n
a
n
d
r
a
n
g
i
n
g
i
n
e
r
t
i
a
l
o
d
o
m
e
t
r
y
,
”
Ad
v
a
n
c
e
d
I
n
t
e
l
l
i
g
e
n
t
S
y
s
t
e
m
s
,
v
o
l
.
5
,
n
o
.
7
,
p
.
2
2
0
0
4
5
9
,
Ju
l
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
2
/
a
i
s
y
.
2
0
2
2
0
0
4
5
9
.
[
5
]
H
.
Y
e
,
Y
.
C
h
e
n
,
a
n
d
M
.
L
i
u
,
“
Ti
g
h
t
l
y
c
o
u
p
l
e
d
3
D
L
i
D
a
r
i
n
e
r
t
i
a
l
o
d
o
me
t
r
y
a
n
d
m
a
p
p
i
n
g
,
”
i
n
2
0
1
9
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
Ro
b
o
t
i
c
s
a
n
d
Au
t
o
m
a
t
i
o
n
(
I
C
RA)
,
I
EEE,
M
a
y
2
0
1
9
,
p
p
.
3
1
4
4
–
3
1
5
0
.
d
o
i
:
1
0
.
1
1
0
9
/
I
C
R
A
.
2
0
1
9
.
8
7
9
3
5
1
1
.
[
6
]
Z.
W
a
n
g
,
L
.
Z
h
a
n
g
,
Y
.
S
h
e
n
,
a
n
d
Y
.
Z
h
o
u
,
“
D
-
LI
O
M
:
t
i
g
h
t
l
y
-
c
o
u
p
l
e
d
d
i
r
e
c
t
Li
D
A
R
-
i
n
e
r
t
i
a
l
o
d
o
me
t
r
y
a
n
d
m
a
p
p
i
n
g
,
”
I
E
E
E
T
ra
n
s
a
c
t
i
o
n
s
o
n
M
u
l
t
i
m
e
d
i
a
,
v
o
l
.
2
5
,
p
p
.
3
9
0
5
–
3
9
2
0
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
T
M
M
.
2
0
2
2
.
3
1
6
8
4
2
3
.
[
7
]
S
.
Zh
a
o
,
Z.
F
a
n
g
,
H
.
L
i
,
a
n
d
S
.
S
c
h
e
r
e
r
,
“
A
r
o
b
u
st
l
a
ser
-
i
n
e
r
t
i
a
l
o
d
o
met
r
y
a
n
d
m
a
p
p
i
n
g
me
t
h
o
d
f
o
r
l
a
r
g
e
-
sc
a
l
e
h
i
g
h
w
a
y
e
n
v
i
r
o
n
m
e
n
t
s,
”
i
n
2
0
1
9
I
E
EE
/
RS
J
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
I
n
t
e
l
l
i
g
e
n
t
R
o
b
o
t
s
a
n
d
S
y
s
t
e
m
s
(
I
RO
S
)
,
I
EEE,
N
o
v
.
2
0
1
9
,
p
p
.
1
2
8
5
–
1
2
9
2
.
d
o
i
:
1
0
.
1
1
0
9
/
I
R
O
S
4
0
8
9
7
.
2
0
1
9
.
8
9
6
7
8
8
0
.
[
8
]
T.
S
h
a
n
,
B
.
En
g
l
o
t
,
D
.
M
e
y
e
r
s,
W
.
W
a
n
g
,
C
.
R
a
t
t
i
,
a
n
d
D
.
R
u
s
,
“
LI
O
-
S
A
M
:
t
i
g
h
t
l
y
-
c
o
u
p
l
e
d
L
i
D
a
r
i
n
e
r
t
i
a
l
o
d
o
m
e
t
r
y
v
i
a
smo
o
t
h
i
n
g
a
n
d
ma
p
p
i
n
g
,
”
i
n
2
0
2
0
I
EEE/
RS
J
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
I
n
t
e
l
l
i
g
e
n
t
Ro
b
o
t
s
a
n
d
S
y
st
e
m
s
(
I
R
O
S
)
,
I
EEE,
O
c
t
.
2
0
2
0
,
p
p
.
5
1
3
5
–
5
1
4
2
.
d
o
i
:
1
0
.
1
1
0
9
/
I
R
O
S
4
5
7
4
3
.
2
0
2
0
.
9
3
4
1
1
7
6
.
[
9
]
J.
Le
v
i
n
s
o
n
e
t
a
l
.
,
“
T
o
w
a
r
d
s
f
u
l
l
y
a
u
t
o
n
o
mo
u
s
d
r
i
v
i
n
g
:
sy
s
t
e
ms
a
n
d
a
l
g
o
r
i
t
h
ms,
”
i
n
2
0
1
1
I
E
EE
I
n
t
e
l
l
i
g
e
n
t
Ve
h
i
c
l
e
s
S
y
m
p
o
s
i
u
m
(
I
V)
,
I
EEE,
Ju
n
.
2
0
1
1
,
p
p
.
1
6
3
–
1
6
8
.
d
o
i
:
1
0
.
1
1
0
9
/
I
V
S
.
2
0
1
1
.
5
9
4
0
5
6
2
.
[
1
0
]
C
.
B
a
d
u
e
e
t
a
l
.
,
“
S
e
l
f
-
d
r
i
v
i
n
g
c
a
r
s
:
a
s
u
r
v
e
y
,
”
E
x
p
e
rt
S
y
st
e
m
s
w
i
t
h
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
6
5
,
p
.
1
1
3
8
1
6
,
M
a
r
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
sw
a
.
2
0
2
0
.
1
1
3
8
1
6
.
[
1
1
]
S
.
L
i
u
e
t
a
l
.
,
“
P
l
a
n
n
i
n
g
d
y
n
a
mi
c
a
l
l
y
f
e
a
si
b
l
e
t
r
a
j
e
c
t
o
r
i
e
s
f
o
r
q
u
a
d
r
o
t
o
r
s
u
s
i
n
g
saf
e
f
l
i
g
h
t
c
o
r
r
i
d
o
r
s
i
n
3
-
D
c
o
mp
l
e
x
e
n
v
i
r
o
n
me
n
t
s
,
”
I
EEE
R
o
b
o
t
i
c
s
a
n
d
Au
t
o
m
a
t
i
o
n
L
e
t
t
e
r
s
,
v
o
l
.
2
,
n
o
.
3
,
p
p
.
1
6
8
8
–
1
6
9
5
,
Ju
l
.
2
0
1
7
,
d
o
i
:
1
0
.
1
1
0
9
/
LR
A
.
2
0
1
7
.
2
6
6
3
5
2
6
.
[
1
2
]
A
.
M
o
k
h
t
a
r
i
,
A
.
B
e
n
a
l
l
e
g
u
e
,
a
n
d
Y
.
O
r
l
o
v
,
“
E
x
a
c
t
l
i
n
e
a
r
i
z
a
t
i
o
n
a
n
d
s
l
i
d
i
n
g
m
o
d
e
o
b
s
e
r
v
e
r
f
o
r
a
q
u
a
d
r
o
t
o
r
u
n
m
a
n
n
e
d
a
e
r
i
a
l
v
e
h
i
c
l
e
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
R
o
b
o
t
i
c
s
a
n
d
A
u
t
o
m
a
t
i
o
n
,
v
o
l
.
2
1
,
n
o
.
1
,
p
p
.
3
9
–
4
9
,
2
0
0
6
,
d
o
i
:
1
0
.
2
3
1
6
/
J
o
u
r
n
a
l
.
2
0
6
.
2
0
0
6
.
1
.
2
0
6
-
2
8
4
2
.
[
1
3
]
S
.
P
a
r
so
n
s,
“
A
u
t
o
n
o
m
o
u
s
r
o
b
o
t
s:
f
r
o
m
b
i
o
l
o
g
i
c
a
l
i
n
s
p
i
r
a
t
i
o
n
t
o
i
mp
l
e
me
n
t
a
t
i
o
n
a
n
d
c
o
n
t
r
o
l
b
y
G
e
o
r
g
e
A
.
B
e
k
e
y
,
M
I
T
p
r
e
ss,
5
6
0
p
p
.
,
$
5
5
.
0
0
,
I
S
B
N
0
-
2
6
2
-
0
2
5
7
8
-
7
,
”
T
h
e
K
n
o
w
l
e
d
g
e
En
g
i
n
e
e
ri
n
g
Re
v
i
e
w
,
v
o
l
.
2
0
,
n
o
.
2
,
p
p
.
1
9
7
–
1
9
8
,
Ju
n
.
2
0
0
5
,
d
o
i
:
1
0
.
1
0
1
7
/
S
0
2
6
9
8
8
8
9
0
5
2
1
0
5
1
2
.
[
1
4
]
G
.
K
i
m
a
n
d
A
.
K
i
m
,
“
S
c
a
n
c
o
n
t
e
x
t
:
e
g
o
c
e
n
t
r
i
c
s
p
a
t
i
a
l
d
e
s
c
r
i
p
t
o
r
f
o
r
p
l
a
c
e
r
e
c
o
g
n
i
t
i
o
n
w
i
t
h
i
n
3
D
p
o
i
n
t
c
l
o
u
d
m
a
p
,
”
i
n
2
0
1
8
I
EEE/
R
S
J
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
I
n
t
e
l
l
i
g
e
n
t
R
o
b
o
t
s
a
n
d
S
y
s
t
e
m
s
(
I
RO
S
)
,
I
EEE,
O
c
t
.
2
0
1
8
,
p
p
.
4
8
0
2
–
4
8
0
9
.
d
o
i
:
1
0
.
1
1
0
9
/
I
R
O
S
.
2
0
1
8
.
8
5
9
3
9
5
3
.
[
1
5
]
G
.
K
i
m,
S
.
C
h
o
i
,
a
n
d
A
.
K
i
m
,
“
S
c
a
n
c
o
n
t
e
x
t
+
+
:
s
t
r
u
c
t
u
r
a
l
p
l
a
c
e
r
e
c
o
g
n
i
t
i
o
n
r
o
b
u
s
t
t
o
r
o
t
a
t
i
o
n
a
n
d
l
a
t
e
r
a
l
v
a
r
i
a
t
i
o
n
s
i
n
u
r
b
a
n
e
n
v
i
r
o
n
m
e
n
t
s,
”
I
EE
E
T
r
a
n
s
a
c
t
i
o
n
s
o
n
Ro
b
o
t
i
c
s
,
v
o
l
.
3
8
,
n
o
.
3
,
p
p
.
1
8
5
6
–
1
8
7
4
,
J
u
n
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
T
R
O
.
2
0
2
1
.
3
1
1
6
4
2
4
.
[
1
6
]
C
.
Y
u
a
n
,
J
.
L
i
n
,
Z.
Li
u
,
H
.
W
e
i
,
X
.
H
o
n
g
,
a
n
d
F
.
Zh
a
n
g
,
“
B
TC
:
a
b
i
n
a
r
y
a
n
d
t
r
i
a
n
g
l
e
c
o
mb
i
n
e
d
d
e
scr
i
p
t
o
r
f
o
r
3
-
D
p
l
a
c
e
r
e
c
o
g
n
i
t
i
o
n
,
”
I
EEE
T
r
a
n
sa
c
t
i
o
n
s
o
n
R
o
b
o
t
i
c
s
,
v
o
l
.
4
0
,
p
p
.
1
5
8
0
–
1
5
9
9
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
T
R
O
.
2
0
2
4
.
3
3
5
3
0
7
6
.
[
1
7
]
C
.
Y
u
a
n
,
J
.
L
i
n
,
Z
.
Z
o
u
,
X
.
H
o
n
g
,
a
n
d
F
.
Z
h
a
n
g
,
“
S
T
D
:
s
t
a
b
l
e
t
r
i
a
n
g
l
e
d
e
s
c
r
i
p
t
o
r
f
o
r
3
D
p
l
a
c
e
r
e
c
o
g
n
i
t
i
o
n
,
”
2
0
2
3
I
E
E
E
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
R
o
b
o
t
i
c
s
a
n
d
A
u
t
o
m
a
t
i
o
n
(
I
C
R
A
)
,
p
p
.
1
8
9
7
–
1
9
0
3
,
M
a
y
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
R
A
4
8
8
9
1
.
2
0
2
3
.
1
0
1
6
0
4
1
3
.
[
1
8
]
M
.
C
u
mm
i
n
s
a
n
d
P
.
N
e
w
ma
n
,
“
F
A
B
-
M
A
P
:
p
r
o
b
a
b
i
l
i
s
t
i
c
l
o
c
a
l
i
z
a
t
i
o
n
a
n
d
map
p
i
n
g
i
n
t
h
e
s
p
a
c
e
o
f
a
p
p
e
a
r
a
n
c
e
,
”
T
h
e
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
Ro
b
o
t
i
c
s
Re
s
e
a
r
c
h
,
v
o
l
.
2
7
,
n
o
.
6
,
p
p
.
6
4
7
–
6
6
5
,
J
u
n
.
2
0
0
8
,
d
o
i
:
1
0
.
1
1
7
7
/
0
2
7
8
3
6
4
9
0
8
0
9
0
9
6
1
.
[
1
9
]
M
.
A
.
F
i
s
c
h
l
e
r
a
n
d
R
.
C
.
B
o
l
l
e
s,
“
R
a
n
d
o
m
sam
p
l
e
c
o
n
se
n
s
u
s:
a
p
a
r
a
d
i
g
m
f
o
r
mo
d
e
l
f
i
t
t
i
n
g
w
i
t
h
a
p
p
l
i
c
a
t
i
o
n
s t
o
i
m
a
g
e
a
n
a
l
y
s
i
s
a
nd
a
u
t
o
m
a
t
e
d
c
a
r
t
o
g
r
a
p
h
y
,
”
C
o
m
m
u
n
i
c
a
t
i
o
n
s
o
f
t
h
e
AC
M
,
v
o
l
.
2
4
,
n
o
.
6
,
p
p
.
3
8
1
–
3
9
5
,
J
u
n
.
1
9
8
1
,
d
o
i
:
1
0
.
1
1
4
5
/
3
5
8
6
6
9
.
3
5
8
6
9
2
.
[
2
0
]
T.
B
o
t
t
e
r
i
l
l
,
S
.
M
i
l
l
s
,
a
n
d
R
.
G
r
e
e
n
,
“
B
a
g
‐
o
f
‐
w
o
r
d
s‐
d
r
i
v
e
n
,
si
n
g
l
e
‐
c
a
m
e
r
a
si
m
u
l
t
a
n
e
o
u
s
l
o
c
a
l
i
z
a
t
i
o
n
a
n
d
ma
p
p
i
n
g
,
”
J
o
u
r
n
a
l
o
f
Fi
e
l
d
R
o
b
o
t
i
c
s
,
v
o
l
.
2
8
,
n
o
.
2
,
p
p
.
2
0
4
–
2
2
6
,
M
a
r
.
2
0
1
1
,
d
o
i
:
1
0
.
1
0
0
2
/
r
o
b
.
2
0
3
6
8
.
[
2
1
]
W
.
X
u
,
Y
.
C
a
i
,
D
.
H
e
,
J.
Li
n
,
a
n
d
F
.
Zh
a
n
g
,
“
F
A
S
T
-
LI
O
2
:
f
a
s
t
d
i
r
e
c
t
Li
D
A
R
-
i
n
e
r
t
i
a
l
o
d
o
me
t
r
y
,
”
I
EE
E
T
r
a
n
s
a
c
t
i
o
n
s
o
n
Ro
b
o
t
i
c
s
,
v
o
l
.
3
8
,
n
o
.
4
,
p
p
.
2
0
5
3
–
2
0
7
3
,
A
u
g
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
T
R
O
.
2
0
2
2
.
3
1
4
1
8
7
6
.
[
2
2
]
A
.
G
e
i
g
e
r
,
P
.
Le
n
z
,
a
n
d
R
.
U
r
t
a
su
n
,
“
A
r
e
w
e
r
e
a
d
y
f
o
r
a
u
t
o
n
o
m
o
u
s
d
r
i
v
i
n
g
?
Th
e
K
I
TTI
v
i
si
o
n
b
e
n
c
h
mar
k
s
u
i
t
e
,
”
i
n
2
0
1
2
I
EE
E
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
u
t
e
r
Vi
s
i
o
n
a
n
d
Pa
t
t
e
r
n
Re
c
o
g
n
i
t
i
o
n
,
I
EEE,
J
u
n
.
2
0
1
2
,
p
p
.
3
3
5
4
–
3
3
6
1
.
d
o
i
:
1
0
.
1
1
0
9
/
C
V
P
R
.
2
0
1
2
.
6
2
4
8
0
7
4
.
[
2
3
]
A
.
G
e
i
g
e
r
,
P
.
L
e
n
z
,
C
.
S
t
i
l
l
e
r
,
a
n
d
R
.
U
r
t
a
su
n
,
“
V
i
si
o
n
mee
t
s
r
o
b
o
t
i
c
s:
t
h
e
K
I
TTI
d
a
t
a
se
t
,
”
T
h
e
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
Ro
b
o
t
i
c
s
Re
se
a
rc
h
,
v
o
l
.
3
2
,
n
o
.
1
1
,
p
p
.
1
2
3
1
–
1
2
3
7
,
S
e
p
.
2
0
1
3
,
d
o
i
:
1
0
.
1
1
7
7
/
0
2
7
8
3
6
4
9
1
3
4
9
1
2
9
7
.
[
2
4
]
R
.
D
u
b
e
,
D
.
D
u
g
a
s,
E.
S
t
u
mm
,
J.
N
i
e
t
o
,
R
.
S
i
e
g
w
a
r
t
,
a
n
d
C
.
C
a
d
e
n
a
,
“
S
e
g
M
a
t
c
h
:
se
g
me
n
t
b
a
se
d
p
l
a
c
e
r
e
c
o
g
n
i
t
i
o
n
i
n
3
D
p
o
i
n
t
c
l
o
u
d
s,
”
i
n
2
0
1
7
I
EE
E
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
R
o
b
o
t
i
c
s
a
n
d
A
u
t
o
m
a
t
i
o
n
(
I
C
RA)
,
I
EEE,
M
a
y
2
0
1
7
,
p
p
.
5
2
6
6
–
5
2
7
2
.
d
o
i
:
1
0
.
1
1
0
9
/
I
C
R
A
.
2
0
1
7
.
7
9
8
9
6
1
8
.
[
2
5
]
C
.
Y
u
a
n
,
W
.
X
u
,
X
.
L
i
u
,
X
.
H
o
n
g
,
a
n
d
F
.
Z
h
a
n
g
,
“
Ef
f
i
c
i
e
n
t
a
n
d
p
r
o
b
a
b
i
l
i
st
i
c
a
d
a
p
t
i
v
e
v
o
x
e
l
m
a
p
p
i
n
g
f
o
r
a
c
c
u
r
a
t
e
o
n
l
i
n
e
L
i
D
A
R
o
d
o
m
e
t
r
y
,
”
I
EEE
Ro
b
o
t
i
c
s
a
n
d
A
u
t
o
m
a
t
i
o
n
L
e
t
t
e
rs
,
v
o
l
.
7
,
n
o
.
3
,
p
p
.
8
5
1
8
–
8
5
2
5
,
J
u
l
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
0
9
/
L
R
A
.
2
0
2
2
.
3
1
8
7
2
5
0
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Lix
ia
o
Ya
n
g
re
c
e
iv
e
d
th
e
B.
E.
d
e
g
re
e
fro
m
Jia
n
g
s
u
Un
i
v
e
rsit
y
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
Zh
e
n
ji
a
n
g
,
Ch
i
n
a
,
i
n
2
0
2
3
.
He
is
c
u
rre
n
tl
y
w
o
rk
i
n
g
t
o
wa
rd
th
e
M
.
S
.
d
e
g
re
e
wi
th
th
e
Oc
e
a
n
Co
ll
e
g
e
,
Jia
n
g
su
U
n
i
v
e
rsity
o
f
S
c
ie
n
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
Z
h
e
n
ji
a
n
g
,
Ch
i
n
a
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
S
LAM
a
n
d
se
n
so
r
fu
si
o
n
.
H
e
c
a
n
b
e
c
o
n
tac
ted
a
t
1
3
0
5
2
7
5
0
0
3
9
@
1
6
3
.
c
o
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.