T
E
L
KO
M
N
I
KA
T
e
lec
om
m
u
n
icat
ion
,
Com
p
u
t
i
n
g,
E
lec
t
r
on
ics
an
d
Cont
r
ol
Vol.
18
,
No.
3
,
J
une
2020
,
pp.
1633
~
1642
I
S
S
N:
1693
-
6930,
a
c
c
r
e
dit
e
d
F
ir
s
t
G
r
a
de
by
Ke
me
nr
is
tekdikti
,
De
c
r
e
e
No:
21/E
/KP
T
/2018
DO
I
:
10.
12928/
T
E
L
KO
M
NI
KA
.
v18i3.
14797
1633
Jou
r
n
al
h
omepage
:
ht
tp:
//
jour
nal.
uad
.
ac
.
id/
index
.
php/T
E
L
K
OM
N
I
K
A
T
h
e
f
lo
w
of
b
ase
li
n
e
e
st
i
m
at
io
n
u
si
n
g a si
n
gl
e
om
n
id
ir
e
c
t
io
n
al
c
a
m
e
r
a
S
u
k
m
a
M
e
gan
ova
E
f
f
e
n
d
i
1
,
Dadet
P
r
am
a
d
ih
an
t
o
2
,
Riyan
t
o
S
igi
t
3
1
D
ep
ar
t
men
t
o
f
Mech
a
t
ro
n
i
c
s
,
Po
l
i
t
e
k
n
i
k
Mek
a
t
ro
n
i
k
a
S
an
at
a
D
h
arma,
In
d
o
n
e
s
i
a
2,
3
E
l
ect
r
o
n
i
cs
E
n
g
i
n
eer
i
n
g
Po
l
y
t
ech
n
i
c
In
s
t
i
t
u
t
e
o
f
Su
ra
b
ay
a
(E
E
PIS)
,
In
d
o
n
es
i
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
Aug
17
,
2019
R
e
vis
e
d
J
a
n
30
,
2020
Ac
c
e
pted
F
e
b
26
,
2020
Bas
el
i
n
e
i
s
a
d
i
s
t
an
ce
b
e
t
w
ee
n
t
w
o
cameras
,
b
u
t
w
e
can
n
o
t
g
et
i
n
fo
rma
t
i
o
n
fro
m
a
s
i
n
g
l
e
camera.
Bas
el
i
n
e
i
s
o
n
e
o
f
t
h
e
i
mp
o
rt
a
n
t
p
arame
t
ers
t
o
fi
n
d
t
h
e
d
ep
t
h
o
f
o
b
j
ec
t
s
i
n
s
t
ere
o
i
mag
e
t
r
i
an
g
u
l
at
i
o
n
.
T
h
e
fl
o
w
o
f
b
a
s
el
i
n
e
i
s
p
ro
d
u
ce
d
b
y
mo
v
i
n
g
t
h
e
camer
a
i
n
h
o
ri
z
o
n
t
al
a
x
i
s
fro
m
i
t
s
o
r
i
g
i
n
a
l
l
o
ca
t
i
o
n
.
U
s
i
n
g
b
as
e
l
i
n
e
es
t
i
mat
i
o
n
,
w
e
can
d
et
erm
i
n
e
d
t
h
e
d
ep
t
h
o
f
an
o
b
j
ect
b
y
u
s
i
n
g
o
n
l
y
a
n
o
m
n
i
d
i
rec
t
i
o
n
a
l
camera.
T
h
i
s
re
s
earch
fo
cu
s
o
n
d
e
t
ermi
n
i
n
g
t
h
e
fl
o
w
o
f
b
as
e
l
i
n
e
b
efo
re
ca
l
cu
l
at
i
n
g
t
h
e
d
i
s
p
ar
i
t
y
map
.
T
o
es
t
i
ma
t
e
t
h
e
fl
o
w
a
n
d
t
o
t
rack
i
n
g
t
h
e
o
b
j
ect
,
w
e
u
s
e
t
h
r
ee
an
d
fo
u
r
p
o
i
n
t
s
i
n
t
h
e
s
u
rface
o
f
an
o
b
j
ec
t
fro
m
t
w
o
d
i
ffere
n
t
d
at
a
(p
an
o
ram
i
c
i
mag
e)
t
h
a
t
w
ere
al
read
y
ch
o
s
en
.
By
mo
v
i
n
g
t
h
e
camera
h
o
r
i
zo
n
t
a
l
l
y
,
w
e
g
et
t
h
e
t
rac
k
s
o
f
t
h
em.
T
h
e
o
b
t
ai
n
e
d
t
rac
k
s
are
v
i
s
u
a
l
l
y
s
i
mi
l
ar.
E
ach
t
rack
re
p
res
e
n
t
t
h
e
co
o
rd
i
n
a
t
e
o
f
each
t
rac
k
i
n
g
p
o
i
n
t
.
T
w
o
o
f
fo
u
r
t
ra
ck
s
h
a
v
e
a
g
rap
h
i
ca
l
rep
res
e
n
t
a
t
i
o
n
s
i
mi
l
ar
t
o
s
eco
n
d
o
rd
er
p
o
l
y
n
o
mi
al
.
K
e
y
w
o
r
d
s
:
B
a
s
e
li
ne
F
low
Omnidi
r
e
c
ti
ona
l
c
a
mer
a
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e
.
C
or
r
e
s
pon
din
g
A
u
th
or
:
S
ukma
M
e
ga
nova
E
f
f
e
ndi
,
De
pa
r
tm
e
nt
of
M
e
c
ha
tr
onics
,
P
oli
teknik
M
e
ka
tr
onika
S
a
na
ta
Dha
r
ma
,
P
a
ingan,
M
a
guwoha
r
jo,
De
pok
,
S
lema
n
55282,
Yo
gya
ka
r
ta,
I
ndone
s
ia.
E
mail:
s
ukma.
mega
nova
@gmail.
c
om
1.
I
NT
RODU
C
T
I
ON
De
pth
of
objec
ts
c
a
n
be
e
s
ti
mate
d
f
r
om
s
ter
e
o
im
a
ge
.
I
n
pr
e
vious
r
e
s
e
a
r
c
h,
s
ter
e
o
im
a
ge
wa
s
obtaine
d
by
us
ing
two
poc
ke
t
c
a
mer
a
s
(
with
im
p
r
ope
r
a
li
gnment
-
ve
r
ti
c
a
l
e
r
r
or
)
[
1]
or
two
omn
idi
r
e
c
ti
ona
l
s
e
ns
or
s
[
2
]
.
Othe
r
r
e
s
e
a
r
c
h
us
e
two
omni
dir
e
c
t
ional
c
a
mer
a
s
(
both
of
them
a
r
e
buil
t
in
ve
r
ti
c
a
l)
[
3
]
,
c
ombi
na
ti
on
of
c
onve
x
m
ir
r
o
r
a
nd
c
onc
a
ve
le
ns
[
4
]
,
two
c
a
mer
a
s
with
two
c
onve
x
mi
r
r
or
s
[
5
,
6
]
,
double
lobed
mi
r
r
or
[
7]
,
a
nd
hype
r
boli
c
double
l
obe
d
mi
r
r
or
[
8]
.
I
n
thi
s
r
e
s
e
a
r
c
h
,
we
only
e
mpl
oy
a
s
ingl
e
omni
dir
e
c
ti
ona
l
c
a
mer
a
with
a
hype
r
boli
c
c
a
tadioptr
ic.
C
a
tadioptr
ic
is
one
of
the
mos
t
im
po
r
tant
c
omp
one
nts
of
a
n
omni
di
r
e
c
ti
ona
l
s
e
ns
or
to
pr
oduc
e
a
n
omni
dir
e
c
ti
ona
l
im
a
ge
.
B
y
us
ing
the
c
a
tadioptr
ic,
a
ll
of
the
objec
ts
a
r
ound
the
omni
di
r
e
c
ti
ona
l
s
e
ns
or
c
a
n
be
s
e
e
n
in
one
omni
dir
e
c
ti
ona
l
im
a
ge
without
a
n
y
e
f
f
or
ts
to
r
otate
the
s
e
ns
or
.
I
t
mea
ns
that
thi
s
s
e
ns
or
ha
s
a
wide
r
f
ield
of
view
:
360°
.
T
he
r
e
a
r
e
many
types
of
c
a
tadioptr
ic
a
nd
e
a
c
h
c
a
tadioptr
ic
will
give
a
dif
f
e
r
e
nt
omni
dir
e
c
ti
ona
l
im
a
ge
.
Ac
c
or
ding
to
[
9
-
15
]
,
t
he
be
s
t
type
o
f
c
a
tadioptr
ic
is
c
a
tadioptr
ic
t
ha
t
ha
s
a
hype
r
boli
c
s
ha
pe
.
T
he
r
e
a
r
e
many
a
dva
ntage
s
us
ing
o
mni
d
ir
e
c
ti
on
a
l
s
e
ns
or
in
c
omput
e
r
vis
ion.
S
ome
of
them
a
r
e
us
ing
f
or
f
a
c
e
de
tec
ti
on
[
16]
,
s
oc
c
e
r
r
obot
(
ba
s
e
d
on
obs
tac
le
de
tec
ti
on)
[
17]
,
matc
hing
a
nd
loca
li
z
a
t
ion
[
18]
.
I
n
the
c
onve
nti
ona
l
s
ter
e
o
im
a
ge
p
r
oc
e
s
s
ing,
ba
s
e
li
ne
is
one
o
f
the
pa
r
a
mete
r
s
to
f
in
d
the
de
pth
of
objec
ts
.
B
y
us
ing
a
pa
ir
o
f
c
a
mer
a
/om
nidi
r
e
c
ti
ona
l
c
a
mer
a
/ca
tadioptr
ics
,
ba
s
e
li
ne
c
a
n
be
de
ter
mi
n
e
d
f
r
o
m
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1633
-
1642
1634
the
dis
tanc
e
be
twe
e
n
c
a
mer
a
s
.
L
ukier
s
ki,
L
e
utene
gge
r
,
a
nd
Da
vis
on
in
[
19
]
us
e
a
n
omni
dir
e
c
ti
ona
l
c
a
mer
a
on
mobi
le
r
obot
.
I
ts
ba
s
e
li
ne
is
de
ter
mi
ne
d
by
the
l
e
ngth
of
the
mobi
le
r
obot
’
s
tr
a
c
k
a
nd
the
length
o
f
it
s
tr
a
c
k
will
a
f
f
e
c
t
the
de
pth
e
s
ti
mation.
I
n
thi
s
r
e
s
e
a
r
c
h
,
ou
r
ba
s
e
li
ne
e
s
ti
mation
is
not
a
f
f
e
c
ted
by
th
e
length
of
the
tr
a
c
k.
W
he
n
we
tr
y
to
obtain
the
de
pt
h
a
nd
map,
we
s
hould
know
a
bout
ba
s
e
li
ne
.
Us
ing
a
n
omni
dir
e
c
ti
ona
l
c
a
mer
a
with
a
hype
r
boli
c
c
a
tadioptr
ic,
ther
e
is
no
inf
or
mation
a
bout
ba
s
e
li
ne
.
De
pth
map
c
a
n
be
de
ve
loped
us
ing
s
t
e
r
e
o
tr
iangula
ti
on
c
a
lcula
ti
on.
Our
pr
opos
e
d
method
only
f
oc
us
e
s
on
de
ter
mi
ning
the
f
low
o
f
ba
s
e
li
ne
.
T
he
f
low
of
ba
s
e
li
ne
in
our
method
is
e
s
ti
mate
d
us
ing
pa
nor
a
mi
c
im
a
ge
be
c
a
us
e
ne
xt
p
r
oc
e
s
s
is
s
ter
e
o
tr
iangula
ti
on
c
a
lcula
ti
on.
T
he
pa
nor
a
mi
c
im
a
ge
is
obtaine
d
by
t
r
a
ns
f
or
med
the
omni
d
ir
e
c
ti
ona
l
im
a
ge
int
o
pa
nor
a
mi
c
im
a
ge
.
B
y
us
ing
s
ter
e
o
tr
iangul
a
ti
on
a
nd
pa
nor
a
mi
c
im
a
ge
,
va
r
iable
(
x
)
s
hould
ha
ve
a
c
ons
tant
va
lue
in
one
a
xis
[
20
]
a
nd
the
de
pth
c
a
n
be
de
ter
mi
ne
d.
T
his
r
e
s
e
a
r
c
h
pr
oduc
e
s
im
a
ge
s
withi
n
dif
f
e
r
e
nt
va
lue
in
two
a
xe
s
(
,
xy
)
.
I
t
s
hows
that
pa
nor
a
mi
c
im
a
ge
,
e
s
pe
c
ially
e
a
c
h
pixel
(
of
the
objec
ts
)
ha
s
a
unique
f
low
that
r
e
p
r
e
s
e
nted
a
c
e
r
tain
pa
tt
e
r
n
.
F
r
om
the
unique
f
low
o
f
e
a
c
h
pixel
in
the
pa
nor
a
mi
c
im
a
ge
,
we
c
a
n
tr
a
c
k
e
a
c
h
pixel
us
ing
L
uc
a
s
-
Ka
n
a
de
’
s
opti
c
a
l
f
low
a
lgor
it
h
m
[
21
-
23
]
.
I
n
or
de
r
to
that
,
we
ha
ve
to
move
the
c
a
mer
a
c
ont
inuous
ly
in
hor
izonta
l
a
xis
f
r
om
it
s
o
r
igi
na
l
loca
ti
on
a
nd
us
e
a
n
objec
t
f
or
de
tec
ti
on
to
p
r
oduc
e
the
de
s
ir
e
d
tr
a
c
ks
.
As
the
r
e
s
ult
,
we
obtain
c
oor
dinate
s
s
hif
ti
ng
o
f
the
objec
t
that
is
s
a
ve
d
int
o
a
f
il
e
a
nd
tr
a
ns
f
or
m
it
int
o
gr
a
ph
s
to
ge
t
ba
s
e
li
ne
e
s
ti
mation.
2.
RE
S
E
AR
CH
M
E
T
HO
D
I
mage
that
is
c
a
ptur
e
d
us
ing
omni
dir
e
c
ti
ona
l
s
e
ns
or
is
c
a
ll
e
d
a
s
omni
d
ir
e
c
ti
ona
l
i
mage
.
I
t
ha
s
a
blac
k
-
c
ir
c
le
in
the
c
e
nter
of
the
im
a
ge
.
Obje
c
t
s
a
r
ound
it
will
be
r
e
f
lec
ted
int
o
the
c
a
tadioptr
i
c
mi
r
r
or
a
nd
the
C
C
D
s
e
n
s
or
f
r
om
the
c
a
mer
a
will
c
a
ptur
e
the
objec
ts
.
T
he
omni
dir
e
c
ti
ona
l
im
a
ge
is
tr
a
ns
f
or
med
int
o
pa
nor
a
mi
c
im
a
ge
be
f
or
e
c
a
li
br
a
ti
on
pr
oc
e
s
s
.
Af
t
e
r
wa
r
d,
we
c
a
n
us
e
the
pa
nor
a
mi
c
im
a
ge
s
to
obtain
f
low
of
ba
s
e
li
ne
e
s
ti
mation.
T
he
s
tr
uc
tur
e
of
omn
idi
r
e
c
ti
ona
l
c
a
mer
a
s
hown
in
F
igur
e
1
.
T
he
mi
r
r
or
or
c
a
tadioptr
ic
s
hould
be
put
a
nd
c
e
nter
e
d
on
the
uppe
r
of
the
c
a
mer
a
.
Unc
e
nter
e
d
mi
r
r
or
wil
l
c
a
us
e
a
dis
tor
ti
on
in
omni
di
r
e
c
ti
on
a
l
im
a
ge
tr
a
ns
f
or
mation
to
pa
nor
a
mi
c
im
a
ge
.
T
o
c
e
nter
the
mi
r
r
or
to
the
c
a
mer
a
,
we
us
e
nuts
,
bol
ts
,
a
nd
s
pr
i
ngs
that
put
in
e
a
c
h
c
or
ne
r
of
the
mi
r
r
or
s
tr
uc
tur
e
.
I
n
thi
s
r
e
s
e
a
r
c
h
,
we
pr
opos
e
a
method
to
e
s
ti
mate
f
low
of
ba
s
e
li
ne
f
r
om
omn
idi
r
e
c
ti
ona
l
im
a
ge
s
.
F
igur
e
1.
S
ys
tem’
s
s
tr
uc
tur
e
of
our
omn
idi
r
e
c
ti
ona
l
c
a
mer
a
us
ing
we
bc
a
m
2.
1.
T
h
e
a
lgori
t
h
m
of
e
s
t
i
m
at
ion
p
r
oc
e
s
s
F
igur
e
2
il
lus
tr
a
tes
a
ll
the
methods
that
us
e
d
in
t
he
f
low
of
ba
s
e
li
ne
e
s
ti
mation
pr
oc
e
s
s
.
T
he
r
e
a
r
e
thr
e
e
(
3)
pr
oc
e
s
s
.
T
he
f
ir
s
t
pr
oc
e
s
s
is
loading
li
ve
im
a
g
e
,
tr
a
ns
f
or
m
omni
di
r
e
c
ti
ona
l
im
a
ge
to
p
a
nor
a
mi
c
im
a
ge
,
a
nd
c
a
li
b
r
a
t
ion
.
S
e
c
ond
pr
oc
e
s
s
is
objec
t
d
e
tec
ti
on
a
nd
tr
a
c
king
the
objec
t’
s
po
int
s
.
T
he
thi
r
d
pr
oc
e
s
s
is
to
plot
the
point
s
of
the
tr
a
c
ks
int
o
gr
a
phs
,
to
f
i
nd
e
qua
ti
on
of
ba
s
e
li
ne
e
s
ti
mation
’
s
f
low,
a
nd
to
d
e
ter
mi
ne
the
c
oe
f
f
icie
nt
o
f
the
e
qua
ti
on
.
2.
2
.
M
od
e
ll
in
g
of
h
yp
e
r
b
oloi
d
m
irr
or
T
h
e
p
r
o
j
e
c
t
i
o
n
f
r
o
m
h
y
p
e
r
b
o
l
i
c
c
a
t
a
d
i
o
p
t
r
i
c
o
r
h
y
p
e
r
b
o
l
o
i
d
m
i
r
r
o
r
t
h
a
t
p
r
o
d
u
c
e
s
a
n
o
m
n
i
d
i
r
e
c
t
i
o
n
a
l
c
a
m
e
r
a
i
s
s
h
o
w
n
i
n
F
i
g
u
r
e
3
.
F
r
o
m
F
i
g
u
r
e
s
3
a
n
d
4
,
t
h
e
e
q
u
a
t
i
o
n
o
f
h
y
p
e
r
b
o
l
o
i
d
m
i
r
r
o
r
m
o
d
e
l
l
i
n
g
i
s
s
h
o
w
n
b
e
l
o
w
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
T
he
fl
ow
of
bas
e
li
ne
e
s
ti
mation
us
ing
a
s
ingl
e
omn
idi
r
e
c
ti
onal
c
ame
r
a
(
Suk
ma
M
e
ganov
a
E
ff
e
ndi
)
1635
(
)
−
−
=
−
2
2
22
1
Zh
R
ab
(
1)
whe
r
e
:
=−
22
a
c
b
.
S
o,
the
map
f
or
hype
r
boloi
d
mi
r
r
o
r
c
a
n
be
c
a
lcula
te
a
s
f
oll
ows
:
=
+
+
t
a
n
Z
R
c
h
(
2)
B
e
c
a
us
e
a
ll
of
the
objec
ts
touch
the
f
ield
a
nd
given
va
lue,
thus
:
+
=−
t
an
ch
R
(
3)
F
r
om
F
igu
r
e
3
,
the
e
qua
ti
on
is
given,
a
s
f
oll
ows
:
(
)
+
=+
−
−
22
22
22
21
t
an
t
an
c
os
b
c
b
c
x
cb
bc
(
4)
=
t
a
n
f
r
(
5)
T
o
s
im
pli
f
y
,
he
nc
e
:
+
=
−
22
1
22
bc
a
bc
;
=
−
2
22
2
bc
a
cb
;
=+
3
()
a
c
h
(
6)
f
oc
a
l
point
of
the
hype
r
boloi
d
mi
r
r
or
is
a
s
s
umed
on
the
p
r
incipa
l
po
int
o
f
the
c
a
mer
a
.
S
o,
(
4
)
to
(
6
)
a
r
e
s
im
pli
f
ied
a
s
f
oll
ows
:
=+
2
1
t
a
n
c
o
s
fa
a
r
(
7)
(
)
(
)
(
)
−
+
=
+
2
22
1
2
12
2
2
t
a
n
t
a
n
1
t
a
n
af
f
aa
r
r
(
8)
B
y
(
3)
a
nd
(
8)
,
the
e
qua
ti
on
is
given
a
s
f
oll
ows
:
−
+
=
+
3
2
2
2
2
1
1
3
3
2
2
2
2
2
2
1
a
af
f
a
a
a
R
a
r
r
R
R
(
9)
(
)
(
)
−
−
+
−
=
2
2
2
2
2
2
2
2
2
1
3
3
1
2
(
)
2
0
f
r
a
R
a
a
f
r
R
r
a
a
a
(
10)
whe
r
e
:
f
=
f
oc
a
l
length
o
f
c
a
mer
a
r
=
r
a
dius
in
pixel
R
=
r
a
dius
in
wor
ld
c
oor
dinate
1
2
3
a
a
a
=
pa
r
a
mete
r
of
hype
r
boloi
d
mi
r
r
or
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1633
-
1642
1636
L
o
a
d
t
h
e
c
a
m
e
r
a
S
t
a
r
t
T
r
a
n
s
f
o
r
m
a
t
i
o
n
o
m
n
i
d
i
r
e
c
t
i
o
n
a
l
i
m
a
g
e
i
n
t
o
p
a
n
o
r
a
m
i
c
i
m
a
g
e
a
n
d
c
a
l
i
b
r
a
t
e
T
r
a
c
k
i
n
g
o
b
j
e
c
t
b
y
c
a
m
e
r
a
m
o
v
e
m
e
n
t
s
S
a
v
e
e
a
c
h
t
r
a
c
k
i
n
g
p
o
i
n
t
s
o
f
o
b
j
e
c
t
i
n
p
a
n
o
r
a
m
i
c
i
m
a
g
e
R
e
p
r
e
s
e
n
t
e
a
c
h
t
r
a
c
k
i
n
g
p
o
i
n
t
s
i
n
t
o
e
a
c
h
g
r
a
p
h
i
c
s
F
i
n
d
t
h
e
e
s
t
i
m
a
t
i
o
n
(
e
q
u
a
t
i
o
n
)
o
f
e
a
c
h
t
r
a
c
k
i
n
g
p
o
i
n
t
s
D
e
t
e
r
m
i
n
e
d
t
h
e
c
o
e
f
f
i
c
i
e
n
t
s
f
r
o
m
t
h
e
e
s
t
i
m
a
t
i
o
n
(
e
q
u
a
t
i
o
n
)
T
h
e
f
l
o
w
o
f
b
a
s
e
l
i
n
e
e
s
t
i
m
a
t
i
o
n
D
e
t
e
c
t
o
b
j
e
c
t
E
n
d
C
a
m
e
r
a
i
s
o
p
e
n
?
N
o
Y
e
s
O
b
j
e
c
t
s
d
e
t
e
c
t
e
d
?
Y
e
s
A
A
N
o
F
igur
e
2
.
T
he
f
low
of
ba
s
e
li
ne
e
s
ti
mation
p
r
oc
e
s
s
α
ᶲ
b
c
c
h
p
(
r
,
θ
)
I
m
a
g
e
p
l
a
n
e
P
(
R
,
Z
,
θ
)
Z
O
m
O
c
R
M
i
r
r
o
r
ᶲ
c
c
h
p
(
r
,
θ
)
I
m
a
g
e
p
l
a
n
e
Z
O
m
O
c
R
W
a
l
l
0
.
5
d
m
e
F
igur
e
3
.
Ge
ometr
y
of
hype
r
boloi
d
c
a
tadioptr
ic/mi
r
r
or
a
nd
p
r
ojec
ti
on
of
point
f
r
om
wor
ld
c
oor
dinate
on
the
im
a
ge
plane
[
24
]
F
igur
e
4
.
I
ns
tallation
s
pe
c
if
ica
ti
on
of
hype
r
boloi
d
c
a
tadioptr
ic/mi
r
r
or
with
a
xis
is
laid
on
the
f
ield
[
24
,
25
]
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
T
he
fl
ow
of
bas
e
li
ne
e
s
ti
mation
us
ing
a
s
ingl
e
omn
idi
r
e
c
ti
onal
c
ame
r
a
(
Suk
ma
M
e
ganov
a
E
ff
e
ndi
)
1637
2.
3
.
Calib
r
at
ion
m
e
t
h
od
W
he
n
the
omni
dir
e
c
ti
ona
l
s
e
ns
or
is
e
mpl
oye
d,
s
ome
dis
tor
ti
on
will
a
ppe
a
r
due
to
it
s
lens
(
im
pr
ope
r
pos
it
ion
)
.
Omnidi
r
e
c
ti
ona
l
s
e
ns
or
pa
r
a
mete
r
s
s
hould
be
a
djus
ted
pr
ope
r
ly
to
mi
nim
ize
the
dis
tor
ti
on.
All
of
the
pa
r
a
mete
r
’
s
a
djus
tm
e
nt
is
c
a
ll
e
d
c
a
li
br
a
ti
on
.
T
he
c
a
li
b
r
a
ti
on
pr
oc
e
s
s
is
one
of
im
po
r
tant
s
tep
be
c
a
us
e
the
omni
di
r
e
c
ti
ona
l
s
e
ns
or
is
us
e
d
to
map
objec
ts
ins
ide
a
r
oom.
T
he
r
e
a
r
e
two
pa
r
a
mete
r
s
in
a
c
a
mer
a
,
e
xt
r
ins
ic
a
nd
int
r
ins
ic.
B
oth
of
them
de
s
c
r
ibe
the
mathe
matica
l
r
e
lations
hip
be
twe
e
n
3D
c
oor
dinate
s
of
r
e
a
l
wo
r
d
s
c
e
n
e
a
nd
the
2D
c
oor
dinate
s
of
it
s
pr
ojec
ti
on
onto
the
im
a
ge
plane
.
S
pe
c
if
ica
ll
y,
the
int
r
ins
ic
or
int
e
r
na
l
pa
r
a
mete
r
is
us
e
d
to
r
e
pr
e
s
e
nt
the
pixel
c
oor
dinate
s
of
im
a
ge
po
int
with
the
c
or
r
e
s
ponding
c
oor
dinate
s
in
the
c
a
mer
a
r
e
f
e
r
e
nc
e
f
r
a
me.
T
he
e
xtr
ins
ic
or
e
xter
na
l
pa
r
a
mete
r
of
the
c
a
mer
a
is
the
pa
r
a
mete
r
that
de
f
ines
the
lo
c
a
ti
on
a
nd
or
ienta
ti
on
of
r
e
a
l
wor
ld
c
oor
dinate
s
t
o
c
a
mer
a
c
oor
dinate
s
.
F
oc
a
l
length
in
pixel
(
,
xy
ff
)
,
pr
incipa
l
point
(
,
xy
cc
)
,
pixel
s
ize
,
r
a
dial
dis
tor
t
ion
(
1
2
3
,,
k
k
k
)
a
nd
tange
nti
a
l
dis
tor
ti
on
(
12
,
pp
)
r
e
pr
e
s
e
nt
the
in
tr
in
s
ic
pa
r
a
mete
r
s
.
T
he
c
a
mer
a
matr
ix
is
r
e
pr
e
s
e
nted
in
3×
3
ma
tr
ix
.
=
0
c
a
m
e
r
a
m
a
t
r
ix
0
0
0
1
xx
yy
fc
fc
(
11)
T
his
is
of
r
a
dial
dis
tor
ti
on.
=
+
+
+
=
+
+
+
2
4
6
1
2
3
2
4
6
1
2
3
d
is
o
r
t
e
d
(
1
)
d
is
o
r
t
e
d
(
1
)
x
y
k
r
k
r
k
r
y
y
k
r
k
r
k
r
(
12)
And
thi
s
is
of
tange
nti
a
l
dis
tor
ti
on
.
(
)
(
)
=
+
+
+
=
+
+
+
22
12
22
12
diso
r
t
e
d
2
2
diso
r
t
e
d
2
2
x
x
p
xy
p
r
x
y
y
p
r
y
p
xy
(
13)
T
he
n,
c
oe
f
f
icie
nt
mus
t
be
known
is
:
(
)
=
1
2
1
2
3
D
i
s
t
o
r
t
i
o
n
c
o
e
f
f
i
c
i
e
n
t
s
k
k
p
p
k
(
14)
M
a
tr
ix
of
r
otation
(
R
)
a
nd
tr
a
ns
lation
(
T
)
a
r
e
r
e
pr
e
s
e
n
ted
be
low.
=
1
1
1
2
1
3
2
1
2
2
2
3
3
1
3
2
3
3
r
r
r
R
r
r
r
r
r
r
(
15)
=
1
2
3
t
Tt
t
(
16)
2
.
4
.
P
a
n
or
am
ic
t
r
an
s
f
or
m
at
ion
Omnidi
r
e
c
ti
ona
l
im
a
ge
is
t
r
a
ns
f
or
med
to
be
a
p
a
nor
a
mi
c
im
a
ge
us
ing
P
ola
r
mapping
.
T
he
P
olar
c
oor
dinate
s
ys
tem
is
mappe
d
int
o
C
a
r
tes
ian
c
oo
r
d
inate
s
ys
tem.
P
o
lar
c
oor
dinate
s
ys
tem
is
de
noted
by
(
,
r
)
a
nd
in
C
a
r
tes
ian
c
oor
dinate
s
ys
tem
is
de
noted
by
(
,
xy
)
.
C
oor
dinate
c
e
nter
of
pola
r
s
a
mpl
ing
is
c
a
lcula
ted
us
ing
(
17)
a
nd
(
18)
.
−
=
1
t
a
n
y
x
(
17)
=+
22
r
x
y
(
18)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1633
-
1642
1638
Va
lue
of
C
a
r
tes
ian
c
oor
dinate
(
,
xy
)
c
a
n
be
c
a
lcula
ted
f
r
om
the
P
ola
r
c
oor
dinate
(
,
r
)
us
ing
tr
igonom
e
tr
y
a
s
f
oll
ows
.
+=
2
2
2
x
y
r
(
19)
=
=
c
o
s
s
i
n
xr
yr
(
20
)
S
o,
the
c
or
r
e
lation
of
C
a
r
tes
ian
c
oor
dinate
s
ys
tem
a
nd
P
olar
c
oor
dinate
s
ys
tem
with
(
,
cc
xy
)
is
the
c
e
nter
poin
t
o
f
the
c
oor
dinate
o
f
omn
idi
r
e
c
ti
ona
l
im
a
ge
(
P
o
lar
c
oor
d
inate
s
ys
tem)
a
nd
a
r
e
given
a
s
f
oll
ows
.
(
)
(
)
=+
=+
0
0
,
c
o
s
,
s
in
c
c
x
r
r
x
y
r
r
y
(
21)
2
.
5
.
Op
t
ical
f
low
I
ntens
it
y
is
one
of
the
pa
r
a
mete
r
s
in
opti
c
a
l
f
low
t
ha
t
us
e
d
f
or
r
e
pr
e
s
e
ntation
the
f
low
o
f
e
ve
r
y
pixel
in
the
im
a
ge
.
B
a
s
e
d
on
it
s
c
ha
r
a
c
ter
is
ti
c
,
th
e
tr
a
c
k
of
e
a
c
h
pixel
c
a
n
be
e
s
ti
mate
d.
W
e
e
mpl
oy
L
uc
a
s
-
Ka
n
a
de
’
s
opti
c
a
l
f
low
a
lgo
r
it
hm.
T
h
is
a
lg
or
it
hm
c
o
mbi
ne
s
s
ome
in
f
or
mation
f
r
om
s
e
ve
r
a
l
ne
a
r
by
pixels
to
r
e
s
olves
the
a
mbi
guit
y
[
21
-
22]
.
T
he
L
uc
a
s
-
Ka
na
de
’s
opti
c
a
l
f
low
a
lgor
it
hm
wor
k
a
s
a
pyr
a
mi
da
l
s
tr
uc
tur
e
withi
n
p
r
oc
e
s
s
f
r
om
the
highes
t
laye
r
unti
l
the
lowe
s
t
laye
r
[
21,
2
3]
.
T
his
a
lgo
r
it
hm
a
s
s
umes
the
tr
a
ns
f
e
r
of
inf
or
mation
in
the
im
a
ge
be
twe
e
n
two
im
a
ge
s
(
f
r
a
mes
)
is
les
s
a
nd
a
lm
os
t
c
on
s
tant
by
c
ons
ider
ing
pixels
ne
a
r
by
of
p
c
oor
dinate
.
Ac
c
or
ding
to
that,
L
uc
a
s
-
Ka
na
de
a
s
s
ume
s
to
maintain
a
ll
pi
xe
ls
with
a
loca
l
windo
w
that
is
c
e
nter
e
d
on
p
c
oor
di
na
te.
T
he
of
it
s
ve
locity
a
r
e
g
iven
a
s
f
oll
ows
.
(
)
(
)
(
)
(
)
(
)
(
)
(
)
(
)
(
)
+
=
−
+
=
−
+
=
−
1
1
1
1
1
1
x
x
y
y
t
x
x
y
y
t
x
n
x
y
n
y
t
n
I
q
V
I
q
V
I
q
I
q
V
I
q
V
I
q
I
q
V
I
q
V
I
q
(
22)
And
(
22)
in
mat
r
ix
A
v
b
=
,
is
r
e
pr
e
s
e
nted
in
(
23)
.
(
)
(
)
(
)
(
)
(
)
(
)
(
)
(
)
(
)
−
−
=
=
=
−
1
11
22
1
,,
t
xy
x
xy
t
y
x
n
y
n
tn
Iq
I
q
I
q
V
I
q
I
q
Iq
A
v
b
V
I
q
I
q
Iq
(
23)
W
he
r
e
:
2
1
,
,
..
..
,
n
q
q
q
=
pixels
in
the
window
(
or
loca
l
window)
(
)
(
)
(
)
,
,
y
x
i
i
t
i
I
q
I
q
I
q
=
pa
r
ti
a
l
de
r
ivatives
o
f
im
a
ge
I
in
a
c
c
or
da
nc
e
with
the
pos
it
ion
(
,
xy
)
a
nd
the
ti
me
(
t
)
on
pixel
i
q
a
t
the
t
im
e
.
L
uc
a
s
-
Ka
n
a
de
’s
a
lgor
it
hm
c
a
n
be
ob
taine
d
us
in
g
lea
s
t
s
qua
r
e
a
lgor
it
hm
a
nd
it
is
c
a
ll
e
d
mat
r
ix
s
ys
tem
of
22
x
TT
A
A
v
A
b
=
o
r
(
)
−
=
1
TT
v
A
A
A
b
(
24)
W
he
r
e
,
T
A
is
tr
a
ns
pos
e
of
matr
ix
A
a
nd
r
e
pr
e
s
e
nted
in
t
he
(
25)
.
(
)
(
)
(
)
(
)
(
)
(
)
(
)
(
)
(
)
(
)
−
−
=
−
1
2
2
x
i
x
i
y
i
x
i
t
i
x
ii
i
y
y
i
t
i
x
i
y
i
y
i
i
ii
I
q
I
q
I
q
I
q
I
q
V
V
I
q
I
q
I
q
I
q
I
q
(
25)
W
it
h
the
tot
a
l
s
umm
ing
a
r
e
f
r
om
i
=
1
unti
l
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
T
he
fl
ow
of
bas
e
li
ne
e
s
ti
mation
us
ing
a
s
ingl
e
omn
idi
r
e
c
ti
onal
c
ame
r
a
(
Suk
ma
M
e
ganov
a
E
ff
e
ndi
)
1639
3.
RE
S
UL
T
S
AN
D
AN
AL
YSI
S
T
he
s
a
mpl
e
o
f
im
a
ge
that
wa
s
c
a
ptur
e
d
us
ing
o
ur
omni
d
ir
e
c
ti
ona
l
c
a
mer
a
is
s
hown
in
F
igur
e
5
.
T
he
omni
di
r
e
c
ti
ona
l
im
a
ge
mus
t
be
tr
a
ns
f
o
r
med
i
nto
pa
nor
a
mi
c
im
a
ge
in
o
r
de
r
to
be
pr
oc
e
s
s
e
d
us
ing
s
ter
e
o
tr
iangula
ti
on.
F
igu
r
e
6
s
hows
the
r
e
s
ult
of
pa
nor
a
mi
c
im
a
ge
that
wa
s
pr
oduc
e
d
us
ing
polar
mapping
a
lgor
it
hm.
T
he
ba
s
e
li
ne
e
s
ti
mation
wa
s
c
a
lcula
ted
us
ing
c
oor
dinate
s
s
hif
ti
ng
of
a
n
objec
t
in
p
a
nor
a
mi
c
im
a
ge
.
C
oor
dinate
s
s
hif
ti
ng
we
r
e
obtaine
d
by
m
oving
the
c
a
mer
a
f
r
om
one
loca
ti
on
to
a
nother
l
oc
a
ti
on.
E
a
c
h
c
oor
dinate
is
a
n
im
a
ge
’
s
pixel
that
wa
s
r
e
pr
e
s
e
nted
by
a
point
(
gr
e
e
n
c
ir
c
le)
in
pa
no
r
a
mi
c
im
a
ge
.
E
a
c
h
point
wa
s
a
s
s
igned
on
the
s
ur
f
a
c
e
of
the
obje
c
t
in
pa
nor
a
mi
c
im
a
ge
.
W
e
a
s
s
ign
th
r
e
e
a
nd
f
ou
r
(
4)
point
s
in
the
de
tec
ted
objec
t
f
or
dif
f
e
r
e
nt
length
of
t
r
a
c
k
(
94
c
m
a
nd
129
c
m)
a
nd
both
of
them
is
s
hown
in
F
igur
e
s
6
(
a
)
a
nd
(
b
)
.
T
he
point
s
we
r
e
us
e
d
a
s
tr
a
c
king
point
s
in
L
uc
a
s
-
Ka
na
de
’
s
opti
c
a
l
f
low
a
l
gor
it
hm
.
F
igur
e
7
s
hows
de
tec
ti
on
point
s
in
the
boa
r
d
we
r
e
s
uc
c
e
s
s
f
ull
y
de
tec
ted
in
thr
e
e
dif
f
e
r
e
nt
im
a
ge
.
T
hos
e
im
a
ge
s
we
r
e
c
a
ptur
e
d
f
r
om
a
dif
f
e
r
e
nt
ti
mes
tamp.
F
igur
e
5
.
Omnidi
r
e
c
ti
ona
l
im
a
ge
(
a
)
(
b)
(
c
)
F
igur
e
6
.
P
a
nor
a
mi
c
im
a
ge
that
wa
s
c
a
ptur
e
d
in
a
r
oom
withi
n
length
of
tr
a
c
k
;
(
a
)
94
c
m
a
nd
(
b)
129
c
m,
(
c
)
P
a
nor
a
mi
c
im
a
ge
tha
t
wa
s
c
a
ptur
e
d
in
labo
r
a
tor
y
(
a
)
(
b)
F
igur
e
7.
F
our
de
tec
ted
point
s
(
a
)
on
length
o
f
t
r
a
c
k
94
c
m
a
nd
thr
e
e
de
tec
ted
point
s
(
b
)
on
length
of
tr
a
c
k
129
c
m
in
the
objec
t
a
nd
the
f
low
o
f
them
f
r
om
the
c
ur
r
e
nt
f
r
a
me
to
the
ne
xt
f
r
a
me
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1633
-
1642
1640
T
he
c
oor
dinate
s
of
tr
a
c
king
point
s
we
r
e
s
a
ve
d
int
o
a
f
il
e
in
*.
txt
f
il
e
f
or
mat
.
T
he
s
a
ve
d
c
oor
dinate
s
in
the
f
il
e
we
r
e
tr
a
ns
f
or
med
int
o
g
r
a
ph
f
o
r
e
a
c
h
point
.
Af
ter
the
g
r
a
phics
we
r
e
obtaine
d
f
or
e
a
c
h
point
,
the
f
low
of
ba
s
e
li
ne
e
s
ti
mation
wa
s
c
a
lcula
ted.
T
he
da
ta
is
tes
ted
in
va
r
ious
c
ondit
ions
ba
s
e
d
on
tr
a
c
k’
s
length
a
nd
ho
r
izonta
l
dis
tanc
e
of
c
a
mer
a
’
s
m
ove
ment.
S
e
ve
r
a
l
da
ta
f
r
om
t
r
a
c
king
point
s
a
r
e
s
hown
in
F
igur
e
s
8
a
nd
9.
T
wo
o
f
the
t
r
a
c
ks
in
both
gr
a
phs
obtain
s
im
il
a
r
f
low
a
nd
the
o
ther
tr
a
c
ks
s
how
a
n
im
pr
ope
r
f
low.
T
he
im
p
r
ope
r
f
low
of
the
tr
a
c
ks
oc
c
ur
s
be
c
a
us
e
of
the
di
f
f
e
r
e
nc
e
l
ight
int
e
ns
it
y
in
the
pa
nor
a
mi
c
im
a
ge
.
T
h
is
phe
nomenon
oc
c
ur
s
be
c
a
us
e
of
the
point
s
de
tec
ti
on
pr
oc
e
s
s
us
e
L
uc
a
s
-
Ka
na
de
’s
opti
c
a
l
f
low
a
lgo
r
it
hm
.
F
igur
e
8.
T
he
tr
a
c
king
f
low
o
f
de
tec
ted
point
s
f
r
o
m
94
c
m
c
a
mer
a
’
s
moveme
nt
F
igur
e
9
.
T
he
tr
a
c
king
f
low
o
f
de
tec
ted
point
s
f
r
o
m
129
c
m
c
a
mer
a
’
s
moveme
nt
A
f
ter
we
c
omput
e
the
e
s
ti
mation
g
r
a
phs
,
the
c
oe
f
f
i
c
ients
of
it
s
a
r
e
g
iven
a
s
f
oll
ows
.
=
=
0
.0
0
1
0
.3
4
9
a
b
T
he
f
low
of
ba
s
e
li
ne
e
s
ti
mation
is
g
iven
a
s
f
oll
ows
:
=
−
+
+
2
0
.
0
0
1
0
.
3
4
9
y
x
x
c
(
22)
Af
ter
we
ge
t
the
ba
s
e
li
ne
e
s
ti
mation,
that
e
qua
ti
on
wa
s
tes
ted
by
t
r
a
c
king
objec
t
in
a
d
if
f
e
r
e
n
t
length
a
nd
dif
f
e
r
e
nt
a
r
e
a
F
igur
e
6
(
c
)
.
T
he
r
e
s
ult
s
hows
the
tr
a
c
k
that
wa
s
pr
oduc
e
d
ha
s
s
im
il
a
r
f
low.
T
he
va
lue
of
c
a
ls
o
dif
f
e
r
e
nt
be
c
a
us
e
it
de
pe
nds
on
the
ini
ti
a
l
pos
it
ion
of
the
tr
a
c
king
point
s
that
a
r
e
not
a
lwa
ys
the
s
a
me.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
T
he
fl
ow
of
bas
e
li
ne
e
s
ti
mation
us
ing
a
s
ingl
e
omn
idi
r
e
c
ti
onal
c
ame
r
a
(
Suk
ma
M
e
ganov
a
E
ff
e
ndi
)
1641
4.
CONC
L
USI
ON
T
he
pr
opos
e
d
method
to
e
s
ti
mate
the
f
low
o
f
ba
s
e
li
ne
f
r
om
a
s
ingl
e
omni
di
r
e
c
ti
ona
l
c
a
mer
a
is
a
c
hieve
d.
F
r
om
the
de
tec
ted
objec
t,
w
e
c
hoos
e
t
hr
e
e
a
nd
f
ou
r
po
int
s
to
tr
a
c
k
a
nd
two
of
the
t
r
a
c
king
point
s
ha
ve
a
s
im
il
a
r
f
low
.
T
he
f
low
of
ba
s
e
li
ne
e
s
ti
mation
method
that
is
obtaine
d
in
thi
s
r
e
s
e
a
r
c
h
s
hows
a
s
e
c
ond
or
de
r
polynom
ial
e
qua
ti
on
with
va
lue
of
0
.
0
0
1
a
=
a
nd
va
lue
of
the
0
.
3
4
9
b
=
.
T
he
c
ons
tant
va
lue
of
c
de
pe
nds
on
the
ini
ti
a
l
pos
it
ion
of
the
t
r
a
c
king
point
s
.
F
r
om
s
e
ve
r
a
l
e
xpe
r
im
e
nts
that
we
r
e
c
onduc
ted,
we
c
onc
l
ude
that
the
e
qua
ti
on
of
ba
s
e
li
ne
e
s
ti
mation
is
not
a
f
f
e
c
te
d
by
length
o
f
tr
a
c
king
point
s
.
I
t
is
indi
c
a
ted
by
the
s
a
me
c
oe
f
f
icie
nt
va
lues
o
f
a
a
nd
b
.
T
he
e
qua
ti
on
of
ba
s
e
li
ne
e
s
ti
mation
that
wa
s
obtaine
d
will
be
us
e
d
f
o
r
the
matc
hing
pr
oc
e
s
s
in
pa
nor
a
mi
c
im
a
ge
.
RE
F
E
RE
NC
E
S
[1
]
J
.
Mro
v
l
j
e
J
an
d
V
.
D
ami
r,
“D
i
s
t
an
ce
Mea
s
u
r
i
n
g
b
as
e
d
o
n
St
ereo
s
co
p
i
c
P
i
ct
u
res
,
”
P
r
o
cee
d
i
n
g
o
f
9
th
In
t
e
r
n
a
t
i
o
n
a
l
P
h
D
W
o
r
ks
h
o
p
o
n
S
ys
t
em
s
a
n
d
Co
n
t
r
o
l
:
Yo
u
n
g
G
en
e
r
a
t
i
o
n
V
i
ew
p
o
i
n
t
,
Izo
l
a,
Sl
o
v
e
n
i
a,
L
j
u
b
l
j
an
a
:
In
s
t
i
t
u
t
J
o
žef
St
efan
,
2
0
0
8
.
[2
]
O
.
M.
M.
K
.
Fo
u
ad
,
“
Po
s
i
t
i
o
n
E
s
t
i
mat
i
o
n
o
f
M
o
b
i
l
e
Ro
b
o
t
s
U
s
i
n
g
O
m
n
i
-
D
i
rec
t
i
o
n
a
l
Cameras
,”
M.
S.
t
h
es
i
s
,
Mech
an
i
cal
E
n
g
i
n
eer
i
n
g
,
Carl
et
o
n
U
n
i
v
er
s
i
t
y
,
O
t
t
a
w
a,
O
n
t
ari
o
,
Can
ad
a,
2
0
1
4
.
[3
]
Z
.
Z
hu
,
“
O
mn
i
d
i
rect
i
o
n
al
St
ere
o
V
i
s
i
o
n
,
”
P
r
o
cee
d
i
n
g
o
f
t
h
e
W
o
r
ks
h
o
p
o
n
O
m
n
i
d
i
r
ect
i
o
n
a
l
V
i
s
i
o
n
,
t
h
e
1
0
t
h
IE
E
E
ICA
R
,
Bu
d
ap
e
s
t
,
H
u
n
g
ar
y
,
2
0
0
1
.
[4
]
S.
Y
i
an
d
N
.
Ahuja
,
“A
n
O
mn
i
d
i
rec
t
i
o
n
a
l
St
ereo
V
i
s
i
o
n
Sy
s
t
em
U
s
i
n
g
a
Si
n
g
l
e
Camera,
”
in
P
r
o
cee
d
i
n
g
o
f
In
t
er
n
a
t
i
o
n
a
l
C
o
n
f
er
e
n
ce
o
n
P
a
t
t
e
r
n
R
ec
o
g
n
i
t
i
o
n
,
p
p
.
8
6
1
-
8
6
5
,
H
o
n
g
K
o
n
g
,
A
u
g
u
s
t
2
0
0
6
.
[5
]
J
.
G
l
u
ck
ma
n
,
S.
K
.
N
ay
ar
,
an
d
K
.
J
.
T
h
o
re
s
z,
“
Real
-
T
i
me
O
mn
i
d
i
rec
t
i
o
n
a
l
an
d
Pan
o
rami
c
S
t
ereo
,
”
i
n
P
r
o
cee
d
i
n
g
s
o
f
t
h
e
1
9
9
8
D
A
R
P
A
Im
a
g
e
U
n
d
e
r
s
t
a
n
d
i
n
g
W
o
r
k
s
h
o
p
,
I
U
W
’
9
8
,
Mo
n
t
ere
y
,
Cal
i
f
o
rn
i
a
,
1
9
9
8
.
[6
]
T
.
Sv
o
b
o
d
a
an
d
T
.
Paj
d
l
a
,
“
E
p
i
p
o
l
ar
G
eo
met
r
y
o
f
Cen
t
ra
l
Cat
a
d
i
o
p
t
ri
c
Camera,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
Co
m
p
u
t
e
r
V
i
s
i
o
n
,
v
o
l
.
49
,
n
o
.
1
,
p
p
.
23
-
3
7
,
A
u
g
u
s
t
2
0
0
2
.
[7
]
D
.
So
u
t
h
w
e
l
l
,
et
a
l
.
,
“Pan
o
ram
i
c
St
ereo
,
”
i
n
P
r
o
c
eed
i
n
g
s
o
f
1
3
t
h
In
t
er
n
a
t
i
o
n
a
l
Co
n
f
e
r
en
ce
o
n
P
a
t
t
e
r
n
R
ec
o
g
n
i
t
i
o
n
,
V
i
e
n
n
a,
A
u
s
t
ri
a,
p
p
.
3
7
8
-
3
8
2
,
1
9
9
6
.
[8
]
E
.
L
.
L
.
Cab
ral
,
J
.
C.
d
e
So
u
za
J
u
n
i
o
r,
an
d
M.
C.
H
u
n
o
l
d
,
“
O
mn
i
d
i
rect
i
o
n
al
S
t
ere
o
V
i
s
i
o
n
w
i
t
h
a
H
y
p
er
b
o
l
i
c
D
o
u
b
l
e
L
o
b
e
d
Mi
rr
o
r,
”
i
n
P
r
o
cee
d
i
n
g
s
o
f
t
h
e
1
7
th
In
t
er
n
a
t
i
o
n
a
l
C
o
n
f
er
e
n
ce
o
n
P
a
t
t
er
n
R
eco
g
n
i
t
i
o
n
,
p
p
.
1
-
9
,
Camb
ri
d
g
e,
U
K
,
A
u
g
u
s
t
2
6
,
2
0
0
4
.
[9
]
T
.
Sat
o
an
d
N
.
Y
o
k
o
y
a
,
“
E
ff
i
ci
e
n
t
H
u
n
d
r
ed
s
-
b
a
s
el
i
n
e
St
ereo
b
y
Co
u
n
t
i
n
g
In
t
er
es
t
Po
i
n
t
s
fo
r
Mo
v
i
n
g
O
mn
i
-
d
i
rect
i
o
n
al
Mu
l
t
i
-
camera
Sy
s
t
em
,”
Jo
u
r
n
a
l
o
f
V
i
s
u
a
l
Co
m
m
u
n
i
ca
t
i
o
n
a
n
d
Im
a
g
e
R
ep
r
es
e
n
t
a
t
i
o
n
,
v
o
l
.
2
1
,
n
o
.
5
-
6
,
p
p
.
4
1
6
-
4
2
6
,
J
u
l
y
-
A
u
g
u
s
t
2
0
1
0
.
[1
0
]
K
.
Y
an
g
Tu
,
“
A
n
a
l
y
s
i
s
an
d
C
o
mp
ar
i
s
o
n
o
f
V
ari
o
u
s
Sh
ap
e
Mi
rr
o
rs
U
s
e
d
i
n
O
mn
i
d
i
rect
i
o
n
al
Camera
s
w
i
t
h
Paramet
er
V
ari
a
t
i
o
n
,
“
Jo
u
r
n
a
l
o
f
In
f
o
r
m
a
t
i
o
n
S
ci
e
n
ce
a
n
d
E
n
g
i
n
eer
i
n
g
,
v
o
l
.
2
5
,
n
o
.
3
,
p
p
9
4
5
-
9
5
9
,
May
2
0
0
9
.
[1
1
]
M.
Y
ach
i
d
a
,
“
O
m
n
i
d
i
rec
t
i
o
n
a
l
Sen
s
i
n
g
a
n
d
Co
m
b
i
n
e
d
Mu
l
t
i
p
l
e
Se
n
s
i
n
g
,
”
i
n
P
r
o
cee
d
i
n
g
s
o
f
IE
E
E
a
n
d
A
TR
W
o
r
ks
h
o
p
o
n
C
o
m
p
u
t
er
V
i
s
i
o
n
f
o
r
V
i
r
t
u
a
l
R
e
a
l
i
t
y
B
a
s
ed
H
u
m
a
n
C
o
m
m
u
n
i
c
a
t
i
o
n
s
,
p
p
.
2
0
-
27
,
B
o
mb
a
y
,
In
d
i
a
,
J
an
u
ary
1
9
9
8
.
[1
2
]
K
.
Y
amazama
K
,
Y
.
Y
ag
i
,
an
d
M.
Y
ac
h
i
d
a,
“
O
mn
i
d
i
rect
i
o
n
al
Ima
g
i
n
g
w
i
t
h
H
y
p
erb
o
l
o
i
d
al
Pr
o
j
ec
t
i
o
n
,”
in
P
r
o
ceed
i
n
g
s
o
f
I
E
E
E
/
R
S
J
In
t
er
n
a
t
i
o
n
a
l
C
o
n
f
er
e
n
ce
o
n
In
t
e
l
l
i
g
e
n
t
R
o
b
o
t
s
a
n
d
S
ys
t
em
s
(IR
O
S
’9
3
)
,
p
p
.
1
0
2
9
-
1
0
3
4
,
Y
o
k
o
h
ama,
J
ap
a
n
,
J
u
l
y
1
9
9
3
.
[1
3
]
J
.
G
as
p
ar,
et
a
l
.
,
“Co
n
s
t
a
n
t
Re
s
o
l
u
t
i
o
n
O
mn
i
d
i
rect
i
o
n
a
l
Cameras
,
”
i
n
P
r
o
cee
d
i
n
g
o
f
O
M
NI
V
IS
'
0
2
W
o
r
ks
h
o
p
o
n
O
m
n
i
-
d
i
r
ect
i
o
n
a
l
V
i
s
i
o
n
,
p
p
.
2
7
-
3
4
,
Co
p
en
h
ag
e
n
,
D
en
m
ark
,
J
u
n
e,
2
0
0
2
.
[
1
4
]
T
.
L
.
C
o
n
r
o
y
a
n
d
J
.
B
.
M
o
o
r
e
,
“
R
e
s
o
l
u
t
i
o
n
I
n
v
a
r
i
a
n
t
S
u
r
f
a
c
e
s
f
o
r
P
a
n
o
r
a
m
i
c
V
i
s
i
o
n
S
y
s
t
e
m
s
,
”
i
n
P
r
o
c
e
e
d
i
n
g
s
o
f
t
h
e
S
e
v
e
n
t
h
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
C
o
m
p
u
t
e
r
V
i
s
i
o
n
,
p
p
.
3
9
2
-
3
9
7
,
K
e
r
k
y
r
a
,
G
r
e
e
c
e
,
S
e
p
t
e
m
b
e
r
2
0
-
2
7
,
1
9
9
9
.
[1
5
]
S.
K
.
N
ay
ar,
“Cat
ad
i
o
p
t
r
i
c
O
m
n
i
d
i
rec
t
i
o
n
a
l
Camera,
”
P
r
o
cee
d
i
n
g
o
f
IE
E
E
Co
m
p
u
t
er
S
o
c
i
et
y
Co
n
f
e
r
en
c
e
o
n
Co
m
p
u
t
er
V
i
s
i
o
n
a
n
d
P
a
t
t
er
n
R
ec
o
g
n
i
t
i
o
n
,
p
p
.
4
8
2
-
4
8
8
,
San
J
u
a
n
,
U
SA
,
J
u
n
e
1
9
9
7
.
[1
6
]
A
.
L
.
C.
Barczak
,
J
.
O
k
amo
t
o
J
r.
,
an
d
V
.
G
ras
s
i
J
r.
,
“Face
T
rack
i
n
g
U
s
i
n
g
a
H
y
p
erb
o
l
i
c
Cat
ad
i
o
p
t
ri
c
O
mn
i
d
i
rect
i
o
n
al
Sy
s
t
em”,
R
es
e
a
r
c
h
Let
t
e
r
s
i
n
t
h
e
I
n
f
o
r
m
a
t
i
o
n
a
n
d
M
a
t
h
em
a
t
i
ca
l
S
ci
e
n
ces
,
v
o
l
.
1
3
,
p
p
.
5
5
-
6
7
,
J
an
u
ary
2
0
0
9
.
[1
7
]
J
.
Si
l
v
a,
et
a
l
.
,
“O
b
s
t
ac
l
e
D
et
ec
t
i
o
n
,
Id
en
t
i
fi
ca
t
i
o
n
an
d
Sh
ari
n
g
o
n
a
Ro
b
o
t
i
c
So
ccer
T
eam,
”
P
r
o
cee
d
i
n
g
o
f
P
o
r
t
u
g
u
e
s
e
Co
n
f
e
r
en
ce
o
n
A
r
t
i
f
i
ci
a
l
In
t
el
l
i
g
en
ce
(
E
P
I
A
2
0
0
9
)
,
p
p
.
3
5
0
-
3
6
0
,
A
v
ei
r
o
,
Po
rt
u
g
a
l
,
2
0
0
9
.
[1
8
]
D
.
Mi
ch
el
,
A
.
A
.
A
rg
y
r
o
s
,
an
d
M.
L
o
u
rak
i
s
,
“H
o
ri
z
o
n
Mat
ch
i
n
g
fo
r
L
o
cal
i
zi
n
g
U
n
o
rd
ere
d
Pan
o
ram
i
c
Imag
e
s
”,
In
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
Co
m
p
u
t
e
r
V
i
s
i
o
n
a
n
d
Im
a
g
e
U
n
d
er
s
t
a
n
d
i
n
g
,
v
o
l
.
1
1
4
,
n
o
.
2
,
p
p
.
2
7
4
-
2
8
5
,
J
an
u
ary
2
0
1
0
.
[1
9
]
R.
L
u
k
i
ers
k
i
,
S.
L
eu
t
en
eg
g
er,
an
d
A
.
J
.
D
av
i
s
o
n
,
“Rap
i
d
Free
-
S
p
ace
Map
p
i
n
g
fro
m
A
Si
n
g
l
e
O
mn
i
d
i
rect
i
o
n
a
l
Camera,
”
i
n
P
r
o
cee
d
i
n
g
o
f
t
h
e
E
u
r
o
p
e
a
n
Co
n
f
e
r
en
ce
o
n
M
o
b
i
l
e
R
o
b
o
t
s
(E
CM
R
)
,
p
p
.
1
-
8
,
L
i
n
co
l
n
,
U
K
,
Sep
t
em
b
er
2
0
1
5
.
[2
0
]
O
p
e
n
CV
D
e
v
T
eam,
“D
ep
t
h
Map
fro
m
St
er
e
o
Imag
e
s
,
”
O
p
e
n
CV
3
.
0
.
0
-
d
e
v
d
o
c
u
men
t
at
i
o
n
:
Camera
Cal
i
b
ra
t
i
o
n
an
d
3
D
Rec
o
n
s
t
r
u
ct
i
o
n
,
N
o
v
1
0
,
2
0
1
4
.
[O
n
l
i
n
e].
A
v
a
i
l
a
b
l
e
fro
m
:
h
t
t
p
s
:
/
/
d
o
c
s
.
o
p
e
n
cv
.
o
rg
/
3
.
0
-
b
et
a
/
d
o
c/
p
y
_
t
u
t
o
ri
al
s
/
p
y
_
cal
i
b
3
d
/
p
y
_
d
ep
t
h
ma
p
/
p
y
_
d
e
p
t
h
map
.
h
t
ml
.
[A
cce
s
s
e
d
N
o
v
em
b
er
9
,
2
0
1
5
].
[2
1
]
R.
Si
g
i
t
,
e
t
a
l
.
,
“Imp
r
o
v
e
d
E
c
h
o
car
d
i
o
g
ra
p
h
y
Seg
me
n
t
a
t
i
o
n
U
s
i
n
g
A
ct
i
v
e
Sh
a
p
e
Mo
d
el
a
n
d
O
p
t
i
cal
F
l
o
w
,
”
In
t
e
r
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
Tel
k
o
m
n
i
k
a
t
el
eco
m
m
u
n
i
c
a
t
i
o
n
co
m
p
u
t
i
n
g
el
ect
r
o
n
i
c
a
n
d
co
n
t
r
o
l
,
v
o
l
.
1
7
,
n
o
.
2
,
p
p
.
8
0
9
-
8
1
8
,
A
p
r
i
l
2
0
1
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
1693
-
6930
T
E
L
KO
M
NI
KA
T
e
lec
omm
un
C
omput
E
l
C
ontr
o
l
,
Vol.
18
,
No
.
3
,
J
une
2020:
1633
-
1642
1642
[2
2
]
A
.
A
.
Prat
i
w
i
,
et
a
l
.
,
“
Imp
ro
v
ed
E
j
ec
t
i
o
n
Fract
i
o
n
Meas
u
reme
n
t
o
n
Card
i
ac
Ima
g
e
U
s
i
n
g
O
p
t
i
ca
l
Fl
o
w
,
”
i
n
P
r
o
ceed
i
n
g
o
f
2
0
1
7
I
n
t
e
r
n
a
t
i
o
n
a
l
E
l
ect
r
o
n
i
c
s
S
ym
p
o
s
i
u
m
o
n
K
n
o
w
l
ed
g
e
C
r
ea
t
i
o
n
a
n
d
I
n
t
e
l
l
i
g
e
n
t
C
o
m
p
u
t
i
n
g
(IE
S
-
KCIC)
IE
E
E
,
p
p
.
2
9
5
-
3
0
0
,
Sep
t
emb
er,
2
0
1
7
.
[2
3
]
U
.
U
mar,
R.
So
el
i
s
t
i
j
o
r
i
n
i
,
an
d
H
.
A
.
D
arw
i
t
o
,
“T
rack
i
n
g
A
ra
h
G
erak
an
T
e
l
u
n
j
u
k
J
ar
i
Berb
as
i
s
W
e
b
c
am
Men
g
g
u
n
ak
a
n
Met
o
d
e
O
p
t
i
ca
l
Fl
o
w
,
”
i
n
P
r
o
cee
d
i
n
g
o
f
t
h
e
1
3
t
h
I
n
d
u
s
t
r
i
a
l
E
l
ec
t
r
o
n
i
cs
S
em
i
n
a
r
2
0
1
1
(I
E
S
2
0
1
1
)
,
p
p
.
2
4
9
-
2
5
4
,
O
ct
o
b
er,
2
0
1
1
.
[2
4
]
M.
J
amzad
,
A
.
R.
H
ad
j
k
h
o
d
a
b
ak
h
s
h
i
,
V
.
S.
Mi
rro
k
n
i
,
“
O
b
j
ect
D
e
t
ect
i
o
n
an
d
L
o
cal
i
za
t
i
o
n
U
s
i
n
g
O
mn
i
d
i
rec
t
i
o
n
a
l
V
i
s
i
o
n
i
n
t
h
e
Ro
b
o
Cu
p
E
n
v
i
r
o
n
me
n
t
,
”
In
t
er
n
a
t
i
o
n
a
l
Jo
u
r
n
a
l
o
f
S
c
i
en
t
i
a
Ir
a
n
i
c
a
,
v
o
l
.
1
4
,
n
o
.
6
,
p
p
.
5
9
9
-
6
1
1
,
N
o
v
emb
er
2
0
0
7
.
[2
5
]
A
.
K
.
Mu
l
y
a,
F.
A
rd
i
l
l
a,
an
d
D
.
Prama
d
i
h
an
t
o
,
“Ba
l
l
T
rack
i
n
g
a
n
d
G
o
a
l
D
et
ec
t
i
o
n
f
o
r
M
i
d
d
l
e
S
i
ze
So
ccer
Ro
b
o
t
U
s
i
n
g
O
mn
i
d
i
rect
i
o
n
al
Camera,
”
P
r
o
ceed
i
n
g
o
f
In
t
er
n
a
t
i
o
n
a
l
E
l
ec
t
r
o
n
i
cs
S
y
m
p
o
s
i
u
m
(IE
S
)
,
p
p
.
4
3
2
-
4
3
7
,
Sep
t
em
b
er
2
0
1
6
.
B
I
OG
RA
P
HI
E
S
OF
AU
T
HO
RS
Sukm
a
M
eg
a
n
o
v
a
Effendi
i
s
c
u
rren
t
l
y
a
mas
t
er
s
t
u
d
e
n
t
w
i
t
h
E
l
ect
r
o
n
i
c
E
n
g
i
n
eer
i
n
g
D
ep
ar
t
men
t
,
E
l
ect
r
o
n
i
cs
E
n
g
i
n
eer
i
n
g
Po
l
y
t
ec
h
n
i
c
In
s
t
i
t
u
t
e
o
f
Su
rab
ay
a
(E
E
PIS),
Su
rab
a
y
a,
In
d
o
n
e
s
i
a.
S
h
e
al
s
o
w
o
rk
s
i
n
Mec
h
at
r
o
n
i
cs
D
ep
ar
t
men
t
,
Po
l
i
t
e
k
n
i
k
Me
k
at
r
o
n
i
k
a
San
a
t
a
D
h
arma,
Y
o
g
y
ak
ar
t
a,
In
d
o
n
e
s
i
a
as
a
l
ect
u
rer.
D
a
det
P
ra
m
a
d
i
ha
nt
o
recei
v
ed
h
i
s
Mas
t
er
d
e
g
ree
i
n
1
9
9
7
a
n
d
P
h
.
D
d
e
g
ree
i
n
2
0
0
3
fr
o
m
D
ep
ar
t
men
t
o
f
Sy
s
t
em
s
an
d
H
u
man
Sc
i
en
ce
s
,
G
rad
u
at
e
Sch
o
o
l
o
f
E
n
g
i
n
eeri
n
g
Sc
i
en
ce,
O
s
a
k
a
U
n
i
v
er
s
i
t
y
,
J
a
p
an
.
H
i
s
res
earc
h
i
n
t
ere
s
t
s
i
n
c
l
u
d
e
co
mp
u
t
er
v
i
s
i
o
n
,
ro
b
o
t
i
c
s
,
real
-
t
i
me
s
y
s
t
ems
,
en
v
i
r
o
n
me
n
t
a
l
mo
n
i
t
o
ri
n
g
,
an
d
cy
b
er
-
p
h
y
s
i
c
al
-
s
y
s
t
ems
.
H
e
i
s
cu
rren
t
l
y
w
o
r
k
i
n
g
as
a
l
ect
u
rer
a
t
Co
mp
u
t
er
E
n
g
i
n
eer
i
n
g
,
D
ep
ar
t
men
t
o
f
I
n
fo
rma
t
i
c
s
an
d
Co
mp
u
t
e
r
E
n
g
i
n
eer
i
n
g
,
an
d
G
rad
u
a
t
e
Sch
o
o
l
o
f
I
n
fo
rma
t
i
c
s
an
d
Co
mp
u
t
er
E
n
g
i
n
eeri
n
g
,
E
l
ect
r
o
n
i
cs
E
n
g
i
n
eeri
n
g
Po
l
y
t
ec
h
n
i
c
In
s
t
i
t
u
t
e
o
f
S
u
rab
a
y
a
(E
E
PIS),
Su
rab
a
y
a,
In
d
o
n
e
s
i
a.
H
e
s
erv
e
s
as
H
ea
d
o
f
E
E
PIS
Ro
b
o
t
i
cs
Cen
t
re
(E
R2
C
-
t
h
e
re
s
earch
cen
t
re
t
h
at
f
o
cu
s
res
earch
o
n
t
h
e
d
ev
e
l
o
p
men
t
o
f
h
u
man
o
i
d
/
h
u
ma
n
l
i
k
e
r
o
b
o
t
s
,
real
-
t
i
me
r
o
b
o
t
i
cs
o
p
era
t
i
n
g
s
y
s
t
em
s
,
co
mp
u
t
er
v
i
s
i
o
n
i
n
r
o
b
o
t
i
c
s
,
an
d
e
n
g
i
n
eer
i
n
g
ed
u
ca
t
i
o
n
b
as
e
d
o
n
ro
b
o
t
i
cs
t
ech
n
o
l
o
g
y
).
R
i
y
a
nto
S
i
g
i
t
recei
v
ed
h
i
s
Ma
s
t
er
d
eg
ree
fro
m
D
ep
ar
t
men
t
o
f
In
f
o
rmat
i
cs
,
In
s
t
i
t
u
t
e
o
f
T
ech
n
o
l
o
g
y
1
0
N
o
v
emb
er
Su
ra
b
ay
a
(I
T
S),
Su
rab
ay
a,
In
d
o
n
e
s
i
a
i
n
1
9
9
5
an
d
Ph
.
D
d
eg
ree
fro
m
U
n
i
v
ers
i
t
i
K
eb
a
n
g
s
aan
Mal
a
y
s
i
a,
Ban
g
i
,
Sel
a
n
g
o
r,
Mal
ay
s
i
a
i
n
2
0
1
4
.
H
e
i
s
cu
rre
n
t
l
y
w
o
r
k
i
n
g
as
a
l
ect
u
rer
at
Co
m
p
u
t
er
E
n
g
i
n
eeri
n
g
,
D
e
p
art
men
t
o
f
In
fo
rma
t
i
c
s
an
d
Co
m
p
u
t
er
E
n
g
i
n
eeri
n
g
,
an
d
G
rad
u
at
e
Sc
h
o
o
l
o
f
I
n
fo
rma
t
i
c
s
an
d
Co
m
p
u
t
er
E
n
g
i
n
eer
i
n
g
,
E
l
ect
r
o
n
i
cs
E
n
g
i
n
eeri
n
g
Po
l
y
t
ech
n
i
c
In
s
t
i
t
u
t
e
o
f
Su
rab
a
y
a
(E
E
PI
S),
Su
rab
ay
a,
In
d
o
n
e
s
i
a.
H
i
s
res
ea
rch
i
n
t
eres
t
s
i
n
cl
u
d
e
me
d
i
ca
l
i
mag
i
n
g
,
art
i
f
i
c
i
al
i
n
t
e
l
l
i
g
e
n
ce,
co
mp
u
t
er
v
i
s
i
o
n
,
an
d
s
e
n
s
o
r.
Evaluation Warning : The document was created with Spire.PDF for Python.