• No se han encontrado resultados

A sensor for vision-based navigation in underwater path tracking with color and edge segmentation

N/A
N/A
Protected

Academic year: 2017

Share "A sensor for vision-based navigation in underwater path tracking with color and edge segmentation"

Copied!
12
0
0

Texto completo

(1)

, ,⋆ , ,

! " # $ %

& '((( ) * + '(((&,

- ./ $ $

" - 0 # 1 2 $ 3

!" #1$3%

$ / 2

2 #

2 4 2 4 2 #

) 2 5 #

4 4 6 #4 $"

7 839 #

4

* : 3 #

2 2 #

6 7

8 2 *#

5

! 9 # 4 5 8

839 *

"

#

$

9 # 4 2

2 4 2 2 4 7

4 2 !

;(+(% 6 2 < #

5 6

4 4 = #

! 4 ;((>% 6 ? #

? 5 #

# 4 # #

2 4

5 # #

!3 3 5 ;((+@ 37 5 . ;(+(%

A A # A ) B A

(2)

, 2 2 2 ? #

2 # 5 4

4

4 2

5 8 #

2 # 4

? 2

2 4 6 ? 2

2 ! ;((+@ 8 ;(++%

6 :

! % 2 2

, ? 2 2 #

2

4 #

# !37 5 . ;(+(%

2 5 ) 4

* C ) 4

4 5 ! % 2 *

< 5

* 4 8 4 2 #

5

, 4 7 6 2

* 2 4 *

? #

4

D 2 *

#? 2

5

%

&

&

, 4 )

5 2 & + ! %

< , *

4 4 4 α xL

! 2 % 2 6

& 2 4

5

4 #5 4

4# 2 # 4 !

4 2 % 4

(3)

4 2 *

/ 5 ) 5

4 4 #

4 *

F 6 #

F * Si 4 F ∪Si

4 Si∩Sj Φ P ! %

* 6 2

P Si $ Si !+%

P Si∪Sj 4 Sj

2 2 HSV #

2 4

* ! 2 % ! %

& + # 3 ? * 19 2

5

% " &

4 4 #

6 83 839 83 ./

#

2 ! 2

(4)

, * Pc .

5 4 ) 5

9 ) * . !.

% 3

Pc Spatch △

$ pij∈Spatch 4 2 !;%

H pij ǫ H −δH, H δH

4 pij 6 Spatch H . 8#2 ! 4 o

o% 839 6 H 8#2

2 δH

2 2

& δH 6

8#2 2

% % &

? 2

8 4 * ;"# *

5 ) !4 %

Pe SlineEstretch △

$ pij ∈SlineEstretch !F%

4 SlineEstretch≡SHough

4 SlineEstretch 6 4 SHough 4

6 8 7

% ' ( $

8 4

? ! 2 ;(('% &

# 4 # 2 4

7 2

8

y −

θ x

r

θ !>%

4 r θ 4 r

θ 2 3

r, θ ? xi, yi

x#y ? r , θ

r#θ

(5)

5

) 6 2

2 7 )

& 6 8 #

2 6 4

2 4 6

r, θ 2 4

2 4 4 6

4 * #

4 2 5

6 5 6

! 6 % * =

& bi bi pj ni ni 2

ni 6 pj 2

bi pj ni→SHough {pj} !G%

i ...Nbi pj i ...Nni→SHough≡SlineEstretch

4 N 4 * 2 2

2

!& 2 ;(('%

'

$ 4 4 2

! % 2 2

& ; 4

& + α xL

2 ! 4 %

) C ./

A H S 2

V 4 7

4 7 ?

2

2 6 #

4 $" 4 7 #

2

)

&

*

/ 2

(6)

* ! 2 % 2 4 4

$" 5

5 2 4

2 4

4 6 2 4

4 ,

2 7

) " + &

& 2 #

xL, α ) 1

! % 2 2

2 3

7 2 2

! & F%

Input frame

Color Separation RGB to HSV

Binariza-tion

Gray scale

Position Estimation

H S V

Binariza-tion

Binariza-tion

Gaussian Filter

Canny edge Detector

Pixel-wise AND Operation

Hough Transform

Shape attributes

Color attributes

Pattern recognition Color patch

& ; 9 # 3

4 < εp

α −α xL −xL ≤δp εp !H%

εp ,

(7)

4 δp #2 δp 4

4 = |α −α | ≤ o |x

L −xL | ≤

6

8 2 bi #

5 r , θ r , θ 2

r −r θ −θ ≤δp εp !I%

εp ,

4 εp ,

δp δp 4

4 = |r −r | ≤ 6 |θ −θ | ≤ o ,

8 2 2 ! %

4

6 G

& F # 7 4 ε

(8)

) % # &

2 < εp

4

* , 4 4

8

2 2 2

* 2 2 2 9 ! 839

% 6 #4 $" 2

4 5 2

2 2 6 7

α xL

4 6

7 V 4

P

V P εp !'%

n

V bi pj 4

SHough {pj} SHough≡SlineEstretch,

4 P . 2 P 6

εp n 6 6

,

6 7 6

V 6 7

*

) ' # ,$

$ 4 2 )

& >

εp εp 2

V ? δV F(J

4

5 ni 2 εp

!I% # 4 ? δV

6 7 ni 4 6

4 ? δV

4 5

6 4 #

(9)

Initial Patch Input frame

Color Separation RGB to HSV

Binariza-tion

Bin checkings

Generation of the performance error

Tracking memory

Threshold Adjustment

H S V

Drop of the external luminance Binariza-tion Binariza-tion I Hough Transform Tangent line estimated

maxni

YES NO p=1 Color Segmentation Algorithm Edge Segmentation Algorithm

& > #

-

.&

$

& 2 6 *

4 ! 5 % 4 <

4 5 4 )

5 4

5 6 5 4 6

4 2 ? 4

2 4

4 5 2

5 6

, 2 4 2

4 6 4 5 ,

2 5

2 4 *6 ! %

2 4

, 4 & G 4 2

εp 4 2 4

& H 4 4

2 2

2 F

4 7 #

2 4 2

(10)

0 20 40 60 80 100 120

failure success

Adaptive Sensor

0 20 40 60 80 100 120

failure success

Fixed-parameter Sensor

frames

& G # 2 2 ! %

*6 # ! %

Line to be tracked

previous line FP

sensor

A sensor

Spurious line

Line to be

tracked previous line

FP sensor

A sensor

Spurious line

Line to be tracked

previous line

A sensor

& H # 9 5 A 4 . A 4

* ! %

* ! * 4% 4 *

4 6 2

εp 4 2

! 4 % 4 *6 #

! % 4 4 4

2 8 *6

4

& I 4 6 #4 $" ! & ;%

(11)

& I # " ) A *6 # !&C%

. A 2 ! %

4 * 4 ) 4

4 4 *

4 4 )

*6 # ? = 4

2 4 4

) *6 # ?

8 4 2 6 =

2 4 4

& 7 2 & '

4

2 + ! % =

2 6 2

9# = 2 2

; 4 !2 %

2 2

F 5 2 ?

4

4 * 4 6

2

/

$

4 5 2 # 2 4 2 #

(12)

5 4 3 2 6 2 4

6 #4 $" 7 #

839 # 4

* : 3

2 2 #

6 7

8 2 *

5

Initial patch Frame 1

Frame 2 Frame 3

& ' # 8 839

A 8# A 3# A 9#

0

+ 2 " !;(('% K. 2 4 8 4 #

8 2 K . .HHF

;((' 1 2

; 8 " @ L 3 , !;((+% K A

2 K C . F> !+;%

F & & 2 !;(('% K. #

2 8 2 K C . 9 >+

+ ;((' C ;MMNF+>

> 8 . !;(++% #

? 33"O;(++ ;(++ +#H

G 8 3 !;(+(% C 5

1 4 9 $ 2 # # 2 / C

2 9 3 13

H 4 8 @ & @ P !;((>% 3 & 3 #

1 6 C #

9 ;H $ M

I 3 / 3 5 / !;((+% Q 9 R ;IM#F;G $ 4

C #8 3 $ (#+F#(F(IMH#F

' 37 5 . !;(+(% 9 A 3 F ;(+(

Referencias

Documento similar

Finally, we adapt also the detection proposal to deal with specific problems of object recognition in different areas. On the one hand, we present a new fully automatic computer

The second method is based on an extension of tensor voting in which the encoding and voting processes are specifically tailored to robust edge detection in color images.. In this

In  this  paper,  the  autonomous  navigation  system  of  the  underwater  vehicle  consisting  of  a  Self‐Organization  Direction  Mapping  Network  (SODMN)  and 

The paper offers two main contributions: a new dynamic envi- ronment model for non-holonomic robots, which maps the environment on the velocity space of the robot; and a set of

The main contributions of the thesis are divided on four aspects: (1) long-term experiments of mobile robot 2D and 3D position tracking in real urban pedestrian scenarios within a

This can be achieved by promoting emotional wellbeing among people diagnosed with long-term conditions, identifying poor mental health quickly and referring to effective help,

 The expansionary monetary policy measures have had a negative impact on net interest margins both via the reduction in interest rates and –less powerfully- the flattening of the

Jointly estimate this entry game with several outcome equations (fees/rates, credit limits) for bank accounts, credit cards and lines of credit. Use simulation methods to