scieee Science in your language
[en] (orig)

Preserving underwater cultural heritage: combining satellite images with a novel flash lidar platform

Author: Bosch, Eleonoor; Langenegger, Fabien; Gege, Peter; Roulet, Jean-Christophe; Nguyen, David; Holzer, Jannis; Cerra, Daniele; Wuethrich, Sonia; Pache, Christophe
Publisher: Zenodo
DOI: 10.4408/IJEGE.2025-01.S-02
Source: https://zenodo.org/records/17296340/files/2_BOSCH.pdf
25
Eleonoo BOSCH(*), Fabien LANGENEGGER(**), Pe e GEGE(***), Jean-Ch is ophe ROULET(*),
Da id NGUYEN(*), Jannis HOLZER(*), Daniele CERRA(***), Sonia WUETHRICH(**) & Ch is ophe PACHE(*)
(*)Sensing & Con ol G oup, CSEM SA - Neuchâ el, Swi ze land
(**)Neuchâ el’s Can onal A chaeology O ice (OARC) - Hau e i e, Swi ze land
(***)Ge man Ae ospace Cen e , Ea h Obse a ion Cen e , Remo e Sensing Technology - Ins i u e Obe p a enho en, Weßling, Deu schland
Co esponding au ho : [email p o ec ed], eleonoo [email p o ec ed]
PRESERVING UNDERWATER CULTURAL HERITAGE:
COMBINING SATELLITE IMAGES WITH A NOVEL FLASH LIDAR PLATFORM
ExTENDED ABSTRACT
Il si o cul u ale di Les A gilliez, nel lago di Neuchâ el in S izze a, a pa e del più ampio si o pa imonio mondiale dell’UNESCO
“Pala i e p eis o iche in o no alle Alpi”. Risale alle cul u e del Co aillod classico (3841-387 a.C.) e del Co aillod a do (ci ca 3500
a.C.). Misu e o og amme iche condo e nel 2014 hanno i ela o la p esenza di 4.834 pala i e in legno in due zone: una occiden ale
densamen e popola a e una o ien ale più spa sa, en ambe a i 2 m e i 3 m di p o ondi à. Nonos an e ques e a ee siano p o e e, il si o
co e due pe icoli: i enomeni di e osione e la p oli e azione di specie in asi e di cozze, nello speci ico la D eissena os i o mis bu-
gensis (cozza quagga) e la D eissena polymo pha (cozza zeb a).
L’e osione, causa a dalle co en i lacus i e dalle a iazioni di p o ondi à, è s a a esace ba a dal iscaldamen o globale, che abbassa
il li ello dell’acqua nei laghi e in luisce sui modelli dei lussi, cos i uendo un pe icolo pe la conse azione del si o pe le gene azioni
u u e. Inol e, la p oli e azione di cozze in asi e po ebbe po a e alla pe di a di ma e iale a cheologico, in pa icola e quello o ganico
come il legno. Pe alu a e e quan i ica e ques i ischi, è essenziale un moni o aggio egola e del si o.
Mappe 3D ad al a isoluzione del si o, se acquisi e egola men e, possono esse e combina e con l’analisi spe ale dei da i delle
immagini sa elli a i pe o ni e in o mazioni comple e sui cambiamen i delle ca a e is iche del ondo del lago e delle s u u e del si o,
nonché sulla p oli e azione delle cozze, in modo da po e alu a e appieno i pe icoli ambien ali del si o. Ques e ecniche po ebbe o
app esen a e un’al e na i a a quelle a uali, che p e edono l’impiego di sommozza o i pe isi a e il si o e documen a e o og a ica-
men e le a iazioni iscon a e. L’analisi au oma izza a e egola e dei da i di poin cloud e delle immagini sa elli a i po ebbe ile a e
empes i amen e enomeni che po ebbe o ichiede e in e en i.
A causa dell’es ensione dell’a ea, della bassa p o ondi à dell’acqua e della a iabili à me eo ologica locale, le ecniche di acquisizione di
immagini subacquee in 3D, come la o og amme ia e il sona , non sono ada e a gene a e equen emen e le mappe 3D necessa ie pe moni-
o a e il si o di Les A gilliez. La o og amme ia é cos osa in e mini di empo in es i o da sommozza o i, p ocessamen o dei da i e iso se
in o ma iche, e la quali à delle immagini dipende dalla aspa enza dell’acqua e dalle condizioni me eo ologiche. I da i sona acquisiscono
ispos e mul iple in acque poco p o onde, che c eano l’e e o mul ipa h deg adando la quali à dei da i. Pe an o, pe la conse azione del si o
e il moni o aggio dell’e osione a Les A gilliez, è s a a s iluppa a una nuo a pia a o ma di acquisizione immagini 3D basa a sulla ecnologia
lash lida (ligh de ec ion and anging), in combinazione con l’analisi spe ale di da i sa elli a i Sen inel-2.
Il lash lida è o imizza o pe le applicazioni subacquee, e p oge a o pe accoglie e da i di poin clouds 3D da un eicolo di su-
pe icie senza equipaggio (USV dall’inglese Unmanned Su ace Vehicle) di medie dimensioni. Il lida è compa o, alimen a o a ba e ia
e do a o di un a ay sul piano ocale di 128×128 pixel. Ciascun pixel è do a o di speci ici ci cui i ele onici e di empo di olo (ToF
dall’inglese Time o Fligh ) pe miglio a e l’acquisizione dei da i in acque dalla ele a a o bidi .
La pia a o ma è s a a p oge a a pe consen i e ile amen i egola i ed e icien i lungo aie o ie p ede ini e, pe me endo un’analisi
coe en e del si o nel empo. Le p ime campagne sono s a e e e ua e lungo linee di e osione p e-p og amma e pe mappa e il p o ilo
del ondo del lago e le sue dinamiche. Un sona mono aggio è s a o u ilizza o in pa allelo al lida pe alida e le misu e di p o ondi à
o enu e. In ine, una se ie di ma oni di e aco a sono s a i posiziona i sul ondo del lago e usa i come i e imen o di dimensioni no e
pe alu a e le p es azioni del lida . In ques o s udio engono p esen a i i isul a i delle campagne di es p elimina i.
Immagini sa elli a i mul ispe ali sono s a e con on a e con gli spe i di i le anza simula i pe sabbia, mac o i e e cozze.
Nell’a icolo engono ipo a i esempi di moni o aggio del si o dallo spazio con immagini del sa elli e Sen inel-2, che iden i icano
anomalie spe ali co ela e alla p oli e azione delle cozze quagga nell’a ea, con e ma e da indagini subacquee. Nuo e immagini Sen-
inel-2 engono acquisi e ogni 5 gio ni, pe me endo un moni o aggio pe iodico dell’a ea pe osse a e le dinamiche del enomeno.
I due ipi di acquisizione di da i p esen a i consen ono di iden i ica e, quan i ica e e accia e in modo e icien e i ischi ambien ali
che minacciano i si i a cheologici subacquei come Les A gilliez.
I alian Jou nal o Enginee ing Geology and En i onmen
www.ijege.uni oma1.i
IJEGE, Special Issue 1 (2025): 25-33, DOI: 10.4408/IJEGE.2025-01.S-02
E-ISSN 2035-5688 | ISSN 1825-6635 / ©Au ho (s)
26 I alian Jou nal o Enginee ing Geology and En i onmen , Special Issue 1 (2025) www.ijege.uni oma1.i
E. BOSCH, F. LANGENEGGER, P. GEGE, J.-C. ROULET, D. NGUYEN, J. HOLZER, D. CERRA, S. WUETHRICH & C. PACHE
ABSTRACT
Les A gilliez, pa o he UNESCO Wo ld He i age Si e
“P ehis o ic Pile Dwellings a ound he Alps,” is in Lake
Neuchâ el (Swi ze land) and da es o he Classical Co aillod
(3841-3817 BC) and La e Co aillod (a ound 3500 BC)
cul u es. Among o he a i ac s, i consis s o 4,834 wooden
piles ound o e a 7000 m2 a ea anging om 2 m o 3 m dep h
below he wa e su ace. Two dange s h ea en he p ese a ion
o he si e: e osion and he p oli e a ion o in asi e mussel
species: D eissena os i o mis bugensis (quagga mussel) and
D eissena polymo pha (zeb a mussel), which pose he speci ic
h ea isk o deg ading he wooden piles.
This s udy p esen s wo me hods de eloped o he
moni o ing o he e osion and mussel popula ions a Les
A gilliez: a new lash lida based 3D imaging pla o m and
sa elli e image da a analysis. Toge he , he wo da a collec ion
schemes allow o e icien iden i ica ion, quan i ica ion
and acking o en i onmen al isks h ea ening unde wa e
a cheological si es, such as his one.
The lash lida is op imized o unde wa e applica ions and
designed o collec 3D poin cloud da a om a medium-sized
unmanned su ace ehicle (USV). The lida sys em is ba e y-
powe ed and ea u es a 128×128 pixel ocal plane a ay o
high esolu ion poin cloud cap u e. When moun ed o he USV,
he lida enables egula and e icien lake su eys, allowing
consis en compa ison o he same loca ions o e ime. The
poin cloud da a enables measu ing he heigh and o ien a ion o
he wooden piles and he lakebed p o ile. The i s demons a i e
lida measu emen esul s a e p esen ed in his s udy.
In addi ion, spec al da a om sa elli e images aken by
Sen inel-2 was compa ed wi h simula ed e lec ance spec a
o sand, mac ophy es and mussels. The p ocessed sa elli e
da a success ully iden i ies spec al anomalies co ela ed o he
p oli e a ion o quagga mussels in he a ea o e a pe iod o ou
yea s, and is con i med by unde wa e su eys done by di e s.
Keywo ds: unde wa e su ey, a chaeology, e osion, lida ,
USV, 3D imaging, emo e sensing, quagga mussels, sen inel-2.
INTRODUCTION
The Les A gilliez cul u al he i age si e is loca ed in Lake
Neuchâ el, Swi ze land (Fig. 1) and is pa o he la ge UNESCO
Wo ld He i age Si e “P ehis o ic Pile Dwellings a ound he
Alps.” I da es o he Classical Co aillod (3841-3817 BC) and
La e Co aillod (a ound 3500 BC) cul u es. A pho og amme ic
su ey conduc ed in 2014 e ealed he p esence o 4,834 wooden
piles o e 7000 m2, some examples a e shown in Fig. 2. The a ea
is di ided in o wo zones: a densely packed wes e n zone whe e
he piles a e ba ely isible, o en ully o pa ially co e ed by he
accumula ion o pebbles, and a mo e sca e ed eas e n zone in
which se e al piles a e exposed up o 50 cm abo e he lakebed
(Wü h ich, 2019). The si e anges om 2 m o 3 m dep h below
he wa e su ace and, despi e being p o ec ed, wo dange s a e
h ea ening i : e osion and he p oli e a ion o in asi e mussel
species: D eissena os i o mis bugensis (quagga mussel) and
D eissena polymo pha (zeb a mussel).
To bo h he eas and wes o he si e, an e osion on is ac i e,
p og essi ely cu ing in o he p o ec ed a ea. The e osion esul s
om he signi ican di e ence in dep h be ween he cen e o he
si e and he pe iphe y, allowing he lake cu en s o p og essi ely
diminish he sedimen deposi s a ound he si e. While less ac i e
han hose obse ed in o he pa s o he lake, his e osion on
mus be cons an ly moni o ed as cu en s in he lake change in
in ensi y om yea o yea and because o he dange o Les
A gilliez. Addi ionally, he p ese a ion o sedimen laye s
is essen ial because hey a e ich in scien i ic in o ma ion.
Obse a ion and analysis o o ganic ma e a Les A gilliez ha e
allowed esea che s o ace he e olu ion o he en i onmen and
ege a ion since he las ice age. Ancien ma shes and housand-
yea -old o es s, p ese ed unde wa e , bea wi ness o changes
in he landscape and he e olu ion o lake le els o e millennia.
The second dange o Les A gilliez is he p oli e a ion o he
in asi e quagga mussel (D eissena os i o mis) which dis up s
ecosys ems and poses a signi ican conce n o a chaeologis s. By
co e ing subme ged si es, hese mussels make i ha de o moni o
p ehis o ic illages, an impo an ask o a chaeological se ices,
and p esen a h ea o u u e gene a ions oppo uni ies o access
and engagemen wi h hem. The p ima y conce n is he po en ial
al e a ion o o ganic emains, including bo h he sedimen laye s
and he piles, by he mussels. Since hei a i al in Lake Neuchâ el,
he e has no been enough ime o assess any possible de e io a ion
Fig. 1 - Ae ial iew o he Les A gilliez cul u al he i age si e on he
No h sho e o Lake Neuchâ el, Swi ze land, wi h map indica -
ing he si e loca ion (inse )
27
I alian Jou nal o Enginee ing Geology and En i onmen , Special Issue 1 (2025) www.ijege.uni oma1.i
PRESERVING UNDERWATER CULTURAL HERITAGE: COMBINING SATELLITE IMAGES WITH A NOVEL FLASH LIDAR PLATFORM
o hese c ucial a chaeological and en i onmen al eco ds.
Two app oaches we e de eloped o moni o ing he h ea s
o Les A gilliez. The mussel popula ions a e acked using da a
p ocessing me hods o emo e sensing da a – namely sa elli e
images om Sen inel-2. In addi ion, he e osion on and mussel
popula ions a e being imaged using a no el lash lida pla o m
which was de eloped o egula ly collec in-si u 3D poin -cloud
da a om he si e. Regula su eys o he e osion on , and pe iodic
moni o ing o he mussel popula ions a e essen ial o unde s and
he e olu ion o he h ea s and, i necessa y, o ake measu es o
he physical p ese a ion o his Wo ld He i age Si e.
Fo analyzing o ganic ma e con en nea he sho e, sa elli e
images aken by Sen inel-2 a e publicly a ailable, and can be
p ocessed wi h spec al analysis echniques which e eal he
c i ical a eas o popula ion g ow h (Ce a, 2024). This echnique
was es ed o he Les A gilliez si e o de e mine whe he he mussel
popula ions could be acked h ough pe iodic analysis o sa elli e
images. The collec ed da a will be combined wi h he esul s o he
3D su ey o con inuous assessmen o mussel g ow h.
Se e al unde wa e cul u al he i age si es use 3D imaging
o moni o and assess he s a e o he si e o o ack he changes
o e ime. Fo example, o he coas o Naples, IT, esea che s
ha e mapped a cul u al he i age si e using pho og amme y and
sona , which enabled biologis s o manually indica e which
3D su aces had some o ganism g ow h based on he images
aken (B uno, 2019). A his si e, wa e cla i y and sunligh
condi ions, as well as he si e dep h, made i possible o use
came as and sona o da a collec ion.
In he Bal ic Sea, a g oup o esea che s es ed a s uc u ed
ligh scanne moun ed o a emo ely ope a ed ehicle (ROV)
which pe o med a 1.5×10 m su ey enabling small objec
econs uc ion wi h 0.8 mm o 1 mm esolu ion (B äue -
Bu cha d , 2023). This sys em does no wo k in shallow
wa e , gi en he size o he ROV alone, bu p o ides excellen
esolu ion o 3D econs uc ion.
Fo gene a ing he 3D poin cloud da a needed in his moni o ing
e o , con en ional unde wa e 3D imaging echniques, such as
pho og amme y and sona we e no sui able. Pho og amme y can
be cos ly in e ms o di e ’s ime as well as compu ing esou ces,
especially when conside ing a 7000 m2 a ea. Fu he , imaging
quali y is dependen on wa e anspa ency and wea he condi ions
and is highly suscep ible o sunligh e ac ion (Calan opio,
2024). Sona imaging, ano he su ey me hod, su e s om
mul iple e lec ions in shallow wa e , causing mul ipa h and
deg ading he quali y o he da a (Hansen, 2011). Fo his, and
many o he cul u al he i age si es, he wa e dep h p ecludes he
use o sona sys ems beyond single beam sona s, which may be
used o co obo a e o he dep h measu emen s.
Unde wa e lida sys ems a e a p omising solu ion o
gene a ing 3D maps e icien ly. Ac i e illumina ion educes
he dependence on wea he condi ions, and no mul ipa h is
gene a ed. Flash lida may employ a 2D single pho on a alanche
de ec o (SPAD) a ay, such as he one in eg a ed in his
esea ch, o collec poin cloud images a high esolu ion and
ame a e. This combina ion enables e icien da a collec ion
o e la ge a eas in all bu high-wind wea he condi ions, u he
enabling epea measu emen s a low inc emen al cos .
This s udy p esen s an o e iew o he sys em deployed o
collec ing 3D poin cloud da a and e iews in de ail he da a and
me hods used in bo h he lida su ey and he sa elli e spec al analysis
o con inuous moni o ing. Finally, esul s o ini ial campaigns a e
p esen ed and discussed ollowed by a conclusion summa izing he
implica ions o he esea ch and a eas o u u e wo k.
LIDAR PLATFORM DESCRIPTION
The egula su ey o Les A gilliez is conduc ed using he no el
lash lida pla o m de eloped o his esea ch. I consis s o a
ba e y-powe ed unmanned su ace ehicle (USV) shown in Fig.
3 equipped wi h na iga ion senso s, he lida , and a single beam
sona . I s dimensions a e abou 128 cm × 80 cm × 60 cm, and
he weigh is 28 kg (wi hou he lida ). I can ei he be manually
emo ely con olled o p og ammed o ollow p e-de ined
ajec o ies using an ine ial measu emen uni (IMU) coupled wi h
a global na iga ion sa elli e sys em (GNSS RTK). The compu e
on he USV uns so wa e o con ol he connec ed equipmen :
1. USV: Ba hyD one (h ps://www.ba hyd one-us .com/)
2. USV au onomous con ol so wa e: QG oundCon ol (h ps://
qg oundcon ol.com)
3. GNSS RTK: P op ie a y i mwa e, da a ans e ed h ough
RS232 o USB-C connec ion.
4. Single beam echo-sounde : Hyd omagic (Eye4so wa e)
5. Lida : CSEM p op ie a y so wa e
QG oundCon ol is an open-sou ce solu ion o PX4 o
A duPilo powe ed ehicles, o e ing comp ehensi e se up and
con ol. I suppo s ull con igu a ion o A duPilo and PX4
Fig. 2 - Subme ged wooden piles in he eas e n zone o Les A gilliez
28 I alian Jou nal o Enginee ing Geology and En i onmen , Special Issue 1 (2025) www.ijege.uni oma1.i
E. BOSCH, F. LANGENEGGER, P. GEGE, J.-C. ROULET, D. NGUYEN, J. HOLZER, D. CERRA, S. WUETHRICH & C. PACHE
P o, acili a es MAVLink p o ocol ope a ions, and includes
mission planning o au onomous missions. Du ing ini ial
es ing, lida da a geo e e encing was done using an ini ial
synch oniza ion o he onboa d sys em, wi h independen
ime-logging o subsequen da a poin s.
The ad an ages o using a USV include minimal en i onmen al
impac due o he absence o a he mic mo o , he abili y o conduc
epea ed missions o imaging o he same loca ions, and a shallow
d a o less han 40 cm, which enables ope a ions in shallow wa e .
The lash lida ( i s p esen ed in Pache, 2021) is moun ed
below he USV, poin ing di ec ly down a he lake loo . The lida
alone is 20×17×19 cm and when encased in wa e p oo housing, i
is 23 cm in diame e and 25 cm all. I is ba e y-powe ed om he
USV wi h a e age powe consump ion below 27 W.
The lida in eg a es a SPAD de ec o used de eloped by he
Fondazione B uno Kessle (I aly), o iginally o space applica ions
(Pe enzoni, 2017). By iling ou de ec o s in a 2×2 a ay, he
sys em has a o al esolu ion o 128×128 pixels. Each pixel is
equipped wi h in eg a ed di ec ime-o - ligh (ToF) measu emen
capabili y and addi ional logic o ope a ion in u bid wa e .
The lida has an adjus able ield o iew (FoV), ocus, and
lase powe o on-si e sys em op imiza ion, depending on
wa e condi ions. Fo su ey a Les A gilliez, a bandpass il e
selec ed o minimize backg ound noise du ing day ime es ing
limi ed he FoV o 35° o a oid igne ing o he lase powe
om he wa eleng h shi h ough he il e .
DATA AND METHODS
To e alua e he changes along he e osion on , ou
measu emen lines pe pendicula o he on we e de ined (Fig.
4). The USV ajec o y was p og ammed o ollow hese lines
while he lida collec s poin cloud da a. Pile leng h om he
lakebed, o hose which can be iden i ied, is measu ed using he
3D poin cloud da a. The measu able leng h o wood piles can
be compa ed be ween es campaigns o assess changes in pile
bu ial dep h which may be due o he mo emen o sedimen .
The poin cloud da a along each line is also a e aged
la e ally o e he FoV o he lida , o ge an a e age dep h
measu emen om he da a collec ed along he de ined
ajec o y. This c ea es one 2D p o ile o lake dep h s. USV
posi ion o each e osion line. The lakebed p o iles allow o
moni o ing he e olu ion o he e osion o e ime.
Te a co a b icks wi h known dimensions we e placed
di ec ly on he lakebed along e osion line B om Fig. 4. These
a ge s a e uniquely iden i iable due o whi e plas ic ma ke s
a ached o each. The b icks can be used o assess he mussel
in asion a e, as hei ma e ial makes hem an ideal subs a e o
mussel a achmen , and hei well-de ined geome y allows o
p ecise e alua ion o mussel de elopmen . The 3D poin clouds
o he b icks and wood piles om he lida , compa ed o e ime,
can p o ide in o ma ion on he mussel popula ion g ow h. Du ing
ini ial measu emen s, he b icks also se ed as e e ence objec s
o e alua e he lida dep h- and spa ial esolu ion pe o mance.
Typical spec a o mussels and o he ma e ials bo h o ganic
and ino ganic con ibu ing o he spec al in o ma ion de ec ed
om space a e analyzed o assess he easibili y o using
sa elli e images o ack mussel popula ions a Les A gilliez
and a which dep hs i could succeed.
Spec a o mix u es o mussels, ypical sand samples
and mac ophy es ha e been simula ed o a dep h o 2.5 m,
conside ing ha he si e is ound a an a e age dep h be ween
2 m and 3 m, using he WASI so wa e (Gege, 2014), and a e
epo ed in Fig. 5. The spec al mix u es ha e been esampled
in he bo om ow o he ele an bands o he Sen inel-2
mul ispec al sa elli e, in o de o assess he impac o hese
h ee componen s o he measu ed wa e spec a. Mussels
Fig. 3 - USV pla o m including lida enclosed in wa e igh housing
( ield o iew acing down)
Fig. 4 - Ae ial iew o he Les A gilliez si e wi h wa e dep h indica ed.
Fou e osion lines, which we e selec ed o dep h p o ile moni-
o ing, a e pic u ed (labelled A-D)
29
I alian Jou nal o Enginee ing Geology and En i onmen , Special Issue 1 (2025) www.ijege.uni oma1.i
PRESERVING UNDERWATER CULTURAL HERITAGE: COMBINING SATELLITE IMAGES WITH A NOVEL FLASH LIDAR PLATFORM
appea o be sepa able om bo h sand and mac ophy es a he
bo om o such shallow inland wa e s, in he g een ange om
500 o 600 nm, co esponding o band 3 in Sen inel-2 da a.
Sa elli e images acqui ed by Sen inel-2 in he p oximi ies o
he si e and he de i ed es ima ed anomaly in he g een band a e
epo ed in he nex sec ion. The ein, anomalies Ag in his ange
o he spec um ha e been compu ed as:
(1)
whe e G1 and G2 a e he median e lec ance in he g een band
o images acqui ed o e he mon h o Ap il in 2018 and 2022,
espec i ely. He e, a no malized di e ence yields a mo e obus
esul , and he h eshold o 0.2 has been adjus ed manually.
Finally, he map ob ained om sa elli e image analysis
was co ela ed wi h di ec obse a ion by scuba di e s om
OARC in au umn 2024.
RESULTS AND DISCUSSION
An example o a econs uc ed scene along e osion line B in
Les A gilliez is shown in Fig. 6 (o ange image). This esul can be
compa ed o an ai bo ne image cap u ed wi h a s anda d came a
moun ed on a small d one (g een image). The a ea cap u ed by
he lida o e he USV ajec o y is delinea ed by he whi e lines
o e layed on he image. The lida image is colo -coded ela i e o
he dep h measu ed by each pixel, wi h ligh e shades indica ing
he measu emen is sho e ( a ge close o he su ace).
The 3D map enables measu emen in all h ee dimensions o
he lakebed ea u es, which is especially impo an in he e alua ion
o he e osion p o ile as discussed in he p e ious sec ion.
In Fig. 7a, he image was cap u ed wi h a s anda d came a and
he pile (ci cled in ed) is ba ely isible due o i s colo blending
wi h he su oundings. In he lida image o he same pile, (Fig.
7b), i is dis inc ly isible, wi h addi ional de ails o i s ip isible
as well as he abili y o de e mine i s heigh om he lakebed.
By compa ing he lida image wi h one aken unde wa e
by a scuba di e (Fig. 7c), he la e al and dep h accu acy and
p ecision o lida measu emen s can be assessed. Based on
he image dis ance scale shown on he igh side o Fig. 7b,
he b ick hickness is es ima ed by he lida o be 14.6 cm, in
e y good ag eemen wi h he nominal b ick hickness o 14
cm. Simila ly, he wood pile, whose heigh om he lakebed
is es ima ed om he poin cloud da a o be 35 cm. F om he
came a image, he pile appea s 2.5 imes alle han he b ick
heigh , indica ing ha he wo imaging me hods ag ee.
The epea abili y o he lida measu emen s is wea he - and
wa e condi ion dependen . Wa es and wind can cause mo ion
blu du ing image cap u e, and algae g ow h and lake cu en s can
Fig. 5 - Simula ions o he impac o mussels and mac ophy es on e lec ance spec a a a wa e dep h o 2.5 m. Top: spec a used in he simula ion
o sand, mussels, and mac ophy es. Bo om: sepa abili y o each pai o spec a o ele an spec al bands in Sen inel 2. Band 3 in he g een
ange shows signi ican disc imina ion be ween mussels and mac ophy es, and bands 1, 2 and 4 p o ide addi ional in o ma ion abou he a io
o mussels and sand o mac ophy es and sand in an image pixel

30 I alian Jou nal o Enginee ing Geology and En i onmen , Special Issue 1 (2025) www.ijege.uni oma1.i
E. BOSCH, F. LANGENEGGER, P. GEGE, J.-C. ROULET, D. NGUYEN, J. HOLZER, D. CERRA, S. WUETHRICH & C. PACHE
cause inc eased u bidi y, educing lida image quali y. The USV
ajec o y is dependen on wea he condi ions as well, as he mo o
con olle mus adap in eal ime o main ain i s heading and speed.
Ini ial lida measu emen s in Les A gilliez s a ed a he end o
2024. Consequen ly, i is oo ea ly o de e mine he p esence and
in ensi y o e osion. The ini ial e osion on p o ile da a om one
o he measu emen lines is shown in Fig. 8. The lida -measu ed
p o ile (in blue) ep esen s he a e age dep h measu ed o e he
la e al FoV (app oxima ely 1.6×1.6 m o a dep h o 3 m).
This a e aging may accoun o disc epancies wi h he
single beam sona p o ile (in o ange), pa icula ly a g ea e
dep hs whe e a di e ence o abou 20 cm is obse ed.
Planned imp o emen s o he lida da a cap u e pla o m,
including be e synch oniza ion ac oss senso s, a e expec ed
o imp o e he lida measu emen pe o mance agains he
sona benchma k.
Fig. 9 p esen s a 3D image o a pile ( ound in e osion line B
om Fig. 4), demons a ing how he dis ance om he pile end
o he lakebed om he lida poin cloud can be used o measu e
e osion o sedimen a ion.
In shallow wa e s nea Concise, app oxima ely 5 km sou h o he
he i age si e o Les A gilliez, esea che s a OARC ha e documen ed
mussel popula ion sp eading. Fig. 11a and Fig. 11b show ende ings
o sa elli e images acqui ed in Ap il 2018 and 2022, espec i ely,
wi h Fig. 11c showing he 2022 image wi h o e laid changes
om 2018 o 2022 de i ed om equa ion (1). The la ges change
Fig. 6 - Compa ison be ween lida acquisi ion (o ange) and a s anda d image aken wi h a came a om an ae ial d one (g een)
Fig. 7 - a) Image om he wa e su ace (no mal came a) o a wooden pile and a es a ge in Les A gilliez. b) Lida dep hmap a he same loca ion.
c) Side- iew o he same es a ge and pile (no mal came a). The es a ge is a e aco a b ick measu ing 25×12×14 cm and he whi e
al eoli ma ke s a e 2×2 cm
31
I alian Jou nal o Enginee ing Geology and En i onmen , Special Issue 1 (2025) www.ijege.uni oma1.i
PRESERVING UNDERWATER CULTURAL HERITAGE: COMBINING SATELLITE IMAGES WITH A NOVEL FLASH LIDAR PLATFORM
is depic ed in ed, con i ming he ex en o he phenomenon and
p o iding a i s iden i ica ion o c i ical mussel p oli e a ion a eas.
Mo ing u he owa ds he subme ged cul u al he i age
si e, a simila dec ease in g een e lec ance was obse ed o e
Sain -Aubin, only 2 km sou h o Les A gilliez and pa ially
co e ing he he i age si e (Fig. 11d, 11e and 11 ). This has been
alida ed on-si e wi h unde wa e su eys (Langenegge ,
2024) which ound quagga mussels in se e al si es, including
bo h Concise and Sain -Aubin. In he Bay o Concise in ensi ied
coloniza ion is epo ed, wi h a eas exploi ed o g a el showing
mussels on he pebbles, con i ming obse a ions om sa elli e
images. A pho og aphic su ey in 2022 ound a clea inc ease
in coloniza ion o quagga mussels in Sain Aubin compa ed o
2019, con i med by he changes de ec ed om sa elli e da a in
Fig. 11d and Fig. 11e, which a e highligh ed in Fig. 11 .
Finally, ae ial images aken wi h a d one in 2022 ha e
signi ican da k a eas compa ed wi h hose o 2018 which
co espond o he pa s colonized by mussels (Fig. 10).
CONCLUSION
The lash lida echnology de eloped by CSEM and he
spec al analysis de eloped a OARC demons a e signi ican
po en ial o moni o ing unde wa e a cheological si es. They
can be e ec i ely used o moni o e osion and in asi e species
popula ions as well as si e deg ada ion.
The compac size, ligh weigh , and low powe consump ion o
he lida make i ideal o in eg a ion on medium-sized Unmanned
Su ace Vehicles (USVs), acili a ing egula and cos -e ec i e
su eys o c i ical si es. Addi ionally, he la e al and measu emen
dis ance accu acy and esolu ion a e well-sui ed o apid and
p ecise ca og aphy o la ge a eas, which can la e be e ined using
mo e p ecise bu cos ly echniques such as pho og amme y.
Open sou ce and inexpensi e sa elli e da a can be used o
ga he in o ma ion on phenomena like he sp ead o in asi e
quagga mussels on he lake bo om. Despi e p omising esul s,
Fig. 8 - Lakebed p o ile measu emen in Les A gilliez (e osion line C).
The blue line is he lida measu ed dep h, and he o ange is
sona measu ed dep h
Fig. 9 - Pile o lakebed heigh measu emen (black line)
Fig. 10 - Inc ease in mussel densi y a he Sain -Aubin (lake Neuchâ el) as seen in came a images aken om a d one
32 I alian Jou nal o Enginee ing Geology and En i onmen , Special Issue 1 (2025) www.ijege.uni oma1.i
E. BOSCH, F. LANGENEGGER, P. GEGE, J.-C. ROULET, D. NGUYEN, J. HOLZER, D. CERRA, S. WUETHRICH & C. PACHE
challenges in spec al sepa abili y pe sis , pa icula ly in
deepe wa e s whe e de ec ion becomes mo e di icul . This
necessi a es u he s udies. Mo eo e , ongoing esea ch
should aim o be e unde s and he implica ions o in asi e
species p oli e a ion o he p ese a ion o unde wa e
a chaeological s uc u es in si es a ec ed by his phenomenon.
The emo e sensing modali ies and examples o deploymen
p esen ed in his pape demons a e ad ances owa d
comp ehensi e unde wa e a cheological si e moni o ing
and isk quan i ica ion. They will be used o egula si e
moni o ing, such ha changes can be iden i ied and acked o
in o m p ese a ion measu es needed o p o ec Les A gilliez.
REFERENCES
B äue -Bu cha d C., Munkel C., Bleie M., Heinze M., Gebha I., Kühms ed P. & No ni G. (2023) - Unde wa e 3d scanning sys em o cul u al
he i age documen a ion. Remo e Sensing, 15(7): 1864. h ps://doi.o g/10.3390/ s15071864ü
B uno F., Lagudi A., Collina M., Medaglia S., Da idde Pe iaggi B., Pe iaggi R., Ricci S. & Sacco Pe asso C. (2019) - Documen a ion and
moni o ing o unde wa e a chaeological si es using 3D imaging echniques: The case s udy o he “Nymphaeum o Pun a Epi a io” (Baiae, Naples).
In e na ional A chi es o he Pho og amme y, Remo e Sensing and Spa ial In o ma ion Sciences, xLII-2/W10: 53-59. h ps://doi.o g/10.5194/isp s-
a chi es-XLII-2-W10-53-2019
Calan opio A. & Chiab ando F. (2024) - Unde wa e Cul u al He i age Documen a ion Using Pho og amme y. Jou nal o Ma ine Science and Engine-
e ing, 12(3): 413. h ps://doi.o g/10.3390/jmse12030413
Ce a D., Gege P., Pla ne S., Langenegge F., Wü h ich S., Roule J.-C. & D oz F. (2024) - Assessing he sp ead o in asi e mussels in lake Neu-
châ el as a po en ial h ea o he Les A gilliez he i age si e wi h Sen inel-2. IGARSS 2024 - 2024 IEEE In e na ional Geoscience and Remo e Sensing
Symposium, A hens, G eece, 2024: 3376-3379. Doi: 10.1109/IGARSS53475.2024.10640529
Gege P. (2014) - WASI-2D: A so wa e ool o egionally op imized analysis o imaging spec ome e da a om deep and shallow wa e s. Compu e s &
Geosciences, 62: 208-215.
Hansen R. E., Callow H. J., Sabo T. O., Synnes S. A. V. (2011) - Challenges in sea loo imaging and mapping wi h syn he ic ape u e sona . IEEE T ansac-
Fig. 11 - (abo e) Sen inel-2 image composi es acqui ed o e Concise (NE) o he mon hs o Ap il 2018 (a) and Ap il 2022 (b) and Sain -Aubin (NE) o
Ap il 2018 (d) and 2022 (e). (below) Anomalies de ec ed in he e olu ion o e lec ance in he g een po ion o he spec um nea Concise be ween
2018 and 2022 (c) and Sain -Aubin be ween 2018 and 2022 ( ), sugges ing coloniza ion o quagga mussels in hese a eas. Values g adually chang-
ing om yellow o ed indica e a ela i e d op in g een e lec ance om 33% o 60% o mo e, espec i ely
33
I alian Jou nal o Enginee ing Geology and En i onmen , Special Issue 1 (2025) www.ijege.uni oma1.i
PRESERVING UNDERWATER CULTURAL HERITAGE: COMBINING SATELLITE IMAGES WITH A NOVEL FLASH LIDAR PLATFORM
ions on Geoscience and Remo e Sensing, 49(10): 3677-3687. h ps://doi.o g/10.1109/TGRS.2011.2155071
Langenegge F. (2024) - Es ima ing he sp ead o in asi e Quagga mussels on he lakebed o lake Neucha el as a po en ial h ea o he Les A gilliez.
OARC in e nal epo , 2024.
Pache C., Meie C., D oz S., Nguyen D., Roule J.-C., Pollini A., Hauge T. F. & D oz F. (2021) - Minia u e Flash lida o Ba hyme y: An Expe i-
men al P oo -o -Concep . 2021 IEEE In e na ional Geoscience and Remo e Sensing Symposium IGARSS, B ussels, Belgium, 2021: 8531-8533.
Doi: 10.1109/IGARSS47720.2021.9554034
Pe enzoni M., Pe enzoni D. & S oppa D. (2017) - A 64×64-Pixels Digi al Silicon Pho omul iplie Di ec TOF Senso Wi h 100-MPho ons/s/pixel Back-
g ound Rejec ion and Imaging/Al ime e Mode Wi h 0.14% P ecision Up To 6 km o Spacec a Na iga ion and Landing. IEEE Jou nal o Solid-S a e
Ci cui s, ol. 52 (1): 151-160, Jan. 2017. Doi: 10.1109/JSSC.2016.2623635
Wü h ich S. (2019) - Go gie - Les A gilliez. Dic ionnai e his o ique de la Suisse (DHS). h ps://hls-dhs-dss.ch/ /a icles/058004/2019-04-03/
Recei ed Feb ua y 2025 - Accep ed July 2025