69
ins umen a ion iewpOin - 21 - MARTECH 23
ID36 PROCESSING ACOUSTIC IMAGES FOR THE EXPLOITATION OF FISH FARMS
F. Gómez-B a o136, A. Ga ocho-C uz137, J. C. Gu ié ez-Es ada138, I. Pulido-Cal o139, A. Pe eg ín-Rubio140, S.
López-Domínguez141, J. Cas o-Gu ié ez142
ABSTRACT
This wo k p esen s a p ocessing me hodology ha uses da a
om a SONAR senso loca ed on a HROV o suppo he exploi a-
ion and managemen o ish a ms.
Keywo ds - Sus ainable aquacul u e, unde wa e Re-
mo ely Ope a ed Vehicles, unde wa e acous ic imaging
INTRODUCTION
The Eu opean Union’s Blue G ow h Plan [1] p oposes ha he u-
u e o ishe y esou ces lies in sma , sus ainable, and inclusi e
aquacul u e p oduc ion, which will play an essen ial ole in p o id-
ing ood o he Eu opean popula ion. To achie e his, i is c ucial
o use di e en echnological solu ions ha suppo he sus ain-
able exploi a ion o aquacul u e a ms and ensu e hei iabili y
[2].
In ecen yea s, mode niza ion e o s in aquacul u e ha e o-
cused on es ima ing g ow h, biomass [3], and physical-chemical
pa ame e s [4]. Addi ionally, unde wa e ehicles equipped wi h
special senso s a e being used o p o ide in o ma ion on he s a e
o he ponds o he a ms [5].
This wo k illus a es a p ocessing me hodology ha allows ex-
ac ing adequa e in o ma ion om he da a p o ided by a MSIS
(Mechanical Scanned Imaging Sona ) senso ha pe o ms a 360º
scan by using an acous ic beam, which has a wide e ical ape -
u e and a na ow ho izon al opening and gene a e acous ic c oss
sec ions o he en i onmen [5].
The p oposed me hodology has been used wi hin he amewo k
o he KTTSeaD ones P ojec , unded by he Eu opean Union,
and lead by he Uni e si y o Huel a. In his p ojec , wo Hyb id
Remo ely Ope a ed Unde wa e Vehicles (HROVs) [5] ha e been
used o suppo he exploi a ion and managemen o ish a ms,
whe e he u bidi y o he wa e is so high ha i is impossible o
use adi ional op ical senso s o isualize he objec s in he en i-
onmen o he ehicle (see Fig 1 a) and Fig 2 a).
Pa icula ly, he objec i e consis s in using acous ic images o lo-
ca ing he elemen s ha a e a ound he HROV and ha may be
o in e es o he exploi a ion o he ins alla ion. To his end, he
p oposed p ocedu e makes i possible o loca e he objec s ela-
i e o he HROV and globally in o de o c ea e a geog aphic coo -
dina es map o he bo om o he pond.
ACOUSTIC IMAGE PROCESSING
The da a acqui ed by he Ping360 sona p o ides in o ma ion
abou he en i onmen o he HROV. Howe e , his in o ma ion
includes edundan da a and noise ha needs o be il e ed o
iden i y po en ial a ge s o in e es o he a m’s ope a o s.
The e o e, a p ocessing mechanism wi h a se ies o s ages has
been implemen ed:
• Fil e ing o he bo de ing a eas wi h he limi o he pond. A
he edges o he pond, he e a e bounces and e lec ions o
he signal ha can lead o he iden i ica ion o alse objec s.
Fo his eason, a mechanism has been de eloped o exclude
he bo de ing a eas wi h he edges o he pond.
• Th eshold il e ing. This mechanism disca ds da a ha a e
he esul o bounces wi h an in ensi y less han o simila o
ha o he signal e u ned by he wa e .
• P oximi y noise il e ing. Da a close o he loca ion o he
HROV a e disca ded due o he appea ance o alse bounces
o s ochas ic na u e.
• Iden i ica ion o elemen s acco ding o he in ensi y o he
e u ned echo and classi ied by p oximi y and connec i i y.
The da a esul ing om il e ing ep esen physical elemen s
ha p o ide an echo signal wi h simila in ensi y and com-
pac spa ial dis ibu ion. The e o e, a p ocessing echnique
based on clus e ing has been applied o g oup hese ele-
men s based on hei p oximi y and connec i i y, as well as
he homogenei y o he in ensi y o he ecei ed signal. As a
esul , a ma ix o elemen s is ob ained, cha ac e ized by he
posi ion o he g oups’ cen oids, hei size, and he in ensi y
o he echo signal.
Once he p ocessing is comple e, he dis ances be ween he HROV
and he cen oid o each elemen a e known, as well as he sec o s
whe e he elemen s a e loca ed. This in o ma ion can be used o
ob ain he posi ion o any o hem e e ed o he local sys em o
he HROV, gi ing he use a ep esen a ion o he solid elemen s
ha su ound he ehicle (see Fig 1 b).
70 ins umen a ion iewpOin - 21 - MARTECH 23
a) b)
Fig 1. a) Na u al ish a m; b) Local Map
Addi ionally, by inco po a ing a global loca ion sys em in o he e-
hicle, i is possible o ca y ou a geog aphic mapping o each o
he iden i ied elemen s. Fig 2 Illus a es an ou doo expe imen ,
whe e se e al objec s we e loca ed in a pool Fig 2 a). Fig 2 b) ep-
esen s he se ing o he expe imen ; Fig 2 c) shows he local map,
and Fig 2 d) p esen s he objec s ep esen ed in a geog aphic co-
o dina es map.
a) c) d)b)
Fig 2. a) Ou doo pool ; b) expe imen se ing; c) local map; d) sa elli e image wi h he de ec ed objec .
CONCLUSIONS
This a icle illus a es a me hodology o p ocessing he in o -
ma ion p o ided by a MSIS senso in o de o de elop local and
global maps o he bo om o a pond in a ish a m. The p o-
posed app oach has been es ed in eal scena ios in he con ex
o he KTTSeaD ones P ojec (‘Conocimien o y ans e encia de
ecnología sob e ehículos aé eos y acuá icos pa a el desa ollo
ans on e izo de ciencias ma inas y pesque as 0622-KTTSEAD-
RONES-5-E’) ounded by he Eu opean Regional De elopmen
Fund (FEDER) h ough he In e eg P og am V-A España-Po ugal
(POCTEP) 2014-2020.
REFERENCES
[1] h ps://ec.eu opa.eu/commission/p essco ne /de ail/en/ip_21_2341, las access 2022/25/09.
[2] Balaban, M.O., So iano, M.G., Ruiz, E.G: Using image analysis o p edic he weigh o Alaskan salmon o di e en species. Jou nal o Food Science 75, 157–162
(2010).
[3] Se pa, D., Fe ei a, P., Fe ei a, H., Fonseca, L.C., Dinis, M.T., Dua e, P.: Modelling he g ow h o whi e seab eam and gil head seab eamin semi-in ensi e ea h
p oduc ion ponds using he Dynamic Ene gy Budge app oach. Jou nal o Sea Resea ch 76: 135–145 (2013).
[4] Gu ié ez-Es ada, J.C., de Ped o, E., López-Luque, R., Pulido-Cal o, I. Compa ison be ween adi ional me hods and a i icial neu al ne wo ks o ammonia con-
cen a ion o ecas ing in an eel (Anguilla anguilla L.) in ensi e ea ing sys em. Aquacul u al Enginee ing 31: 183-203 (2004).
[5] Gómez-B a o, F., Ga ocho-C uz, A., Gu ié ez-Es ada, J. C., Pulido-Cal o, I., Cas o-Gu ié ez, J., Pe eg ín-Rubio, A., & López-Domínguez, S. (2021). Loca ing i ems
o in e es by using emo ely ope a ed ehicles o he sus ainable exploi a ion o ish a ms. In XLII Jo nadas de Au omá ica (pp. 617-624). Uni e sidade da Co uña,
Se izo de Publicacións.
