The EGIM, a generic instrumental module to equip EMSO observatories

Lantéri, Nadine; Legrand, Julien; Moreau, Bertrand; Lagadec, Jean Romain; Rolin, Jean François

doi:10.1109/oceanse.2017.8084824

Cited by 6 publications

(3 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…During the last two decades, underwater imaging assets have increased with the improvement of High-Definition (HD) optics and the introduction of low-light equipment that, in association with acoustic multi-beam cameras, are presently allowing vision in darkness [3]. Complex sensor packages are being installed on cabled observatories and their docked mobile platforms such as crawlers (i.e., Internet Operated Vehicles, IOVs), combining oceanographic and geo-chemical assets with cameras [20,20,21,21], [32,50,93]. This combined image and multiparametric environmental data acquisition is allowing to link visual counts of animals for the different species within a marine community with concomitant habitat quality changes, as an experimental field measure of their ecological niche [1,2].…”

Section: Marine Imaging Acquisitionmentioning

confidence: 99%

Deep learning based deep-sea automatic image enhancement and animal species classification

et al. 2023

View full text Add to dashboard Cite

The automatic classification of marine species based on images is a challenging task for which multiple solutions have been increasingly provided in the past two decades. Oceans are complex ecosystems, difficult to access, and often the images obtained are of low quality. In such cases, animal classification becomes tedious. Therefore, it is often necessary to apply enhancement or pre-processing techniques to the images, before applying classification algorithms. In this work, we propose an image enhancement and classification pipeline that allows automated processing of images from benthic moving platforms. Deep-sea (870 m depth) fauna was targeted in footage taken by the crawler “Wally” (an Internet Operated Vehicle), within the Ocean Network Canada (ONC) area of Barkley Canyon (Vancouver, BC; Canada). The image enhancement process consists mainly of a convolutional residual network, capable of generating enhanced images from a set of raw images. The images generated by the trained convolutional residual network obtained high values in metrics for underwater imagery assessment such as UIQM (~ 2.585) and UCIQE (2.406). The highest SSIM and PSNR values were also obtained when compared to the original dataset. The entire process has shown good classification results on an independent test data set, with an accuracy value of 66.44% and an Area Under the ROC Curve (AUROC) value of 82.91%, which were subsequently improved to 79.44% and 88.64% for accuracy and AUROC respectively. These results obtained with the enhanced images are quite promising and superior to those obtained with the non-enhanced datasets, paving the strategy for the on-board real-time processing of crawler imaging, and outperforming those published in previous papers.

show abstract

Section: Marine Imaging Acquisitionmentioning

confidence: 99%

Deep learning based deep-sea automatic image enhancement and animal species classification

et al. 2023

View full text Add to dashboard Cite

show abstract

“…In order to illustrate this capability, recordings from a hydrophone deployed within an EGIM (EMSO Generic Instrument Module) have been processed [46]. The deployment was performed at the Plataforma Oceánica de Canarias (PLOCAN) in the Canary Islands, Spain, during June and July 2019.…”

Section: B Egim Passive Acoustics Recordingsmentioning

confidence: 99%

Metadata-Driven Universal Real-Time Ocean Sound Measurement Architecture

et al. 2021

View full text Add to dashboard Cite

Underwater sound in the oceans has been significantly rising in the past decades due to an increase in human activities, adversely affecting the marine environment. In order to assess and limit the impact of underwater noise, the European Commission's Marine Strategy Framework Directive (MSFD) included the long-term monitoring of low-frequency underwater sound as a relevant indicator to achieve a good environmental status. There is a wide range of commercial hydrophones and observing platforms able to perform such measurements. However, heterogeneity and lack of standardization in both hydrophones and observing platforms makes the integration and data management tasks time-consuming and error-prone. Moreover, their power and communications constraints need to be addressed to make them suitable for long-term ocean sound monitoring. Measured underwater sound levels are challenging to compare because different measurement methodologies are used, leading to a risk of misunderstandings and data misinterpretation. Furthermore, the exact methodology applied is not always public or accessible, significantly reducing ocean sound data re-usability. Within this work, a universal architecture for ocean sound measurement is presented, addressing hydrophone integration, real-time in situ processing and data management challenges. Emphasis is placed on generic and re-usable components, so it can be seamlessly replicated and deployed in new scenarios regardless of the underlying hardware and software constraints (hydrophone model, observing platform, operating system, etc.). Within the proposed architecture, a generic implementation of an underwater sound algorithm based on underwater noise measurement best practices is provided. Standardized and coherent metadata with emphasis on strong semantics is discussed, providing the building blocks for FAIR (Findable, Accessible, Interoperable, Reusable) ocean sound data management.

show abstract

“…The metadata model to abstract sensors, the interoperable architecture for sensor integration discussed in chapter 2 and the ocean sound algorithm discussed in chapter [75]. It has been designed bearing in mind the heterogeneity of the regional facilities that conform EMSO ERIC [76].…”

Section: Use Casesmentioning

confidence: 99%

Marine sensor interoperability applied to in situ ocean sound processing

Martínez Padró

View full text Add to dashboard Cite

There has always been background sound in the oceans due to natural factors. However, since the first hydrophone deployments at mid-20th century, measured ocean sound levels have been steadily rising due to an increment of human activities such as shipping, construction, sonar and seismic exploration. Ocean sound is an important environmental factor for many species, especially to those using underwater sounds for localization and communication (in example marine mammals). The introduction of energy into the marine habitat in the form of underwater noise is harmful for the ecosystem. Therefore, it needs to be properly monitored to achieve a good environmental status. The intrinsic nature of the marine environment complicates the acquisition of representative underwater sound datasets. Wide ocean extensions at different depths need to be monitored, presenting severe limitations on the available sampling methods. The deployment of fixed stations such as anchored autonomous recorders or moored hydrophone arrays are expensive and present numerous issues such as storage capacity, limited autonomy and low-bandwidth communications. On the other hand, state-of-the-art unmanned vehicles, such as underwater gliders or autonomous surface vehicles have an exceptional endurance, as well as excellent sampling capability over vast areas. However, these observation platforms are even more constrained in terms of power availability and communication's bandwidth. Commercial off-the-shelf hydrophones used in ocean sound monitoring provide raw acoustic data, whether as acoustic recordings or as data streams. Due to the high volume of raw data produced by hydrophones, streaming this data to a shore station using constrained communications is generally not possible. Moreover, the telemetry is usually one of the most power-demanding systems in observation platforms. Thus, in order to obtain ocean sound levels, acoustic data has to be stored on-board and processed offline after the platform's recovery. However, if raw acoustic data could be processed in situ, the volume of data to be sent would be reduced by several orders of magnitude. Therefore, processing acoustic data in situ would allow real-time ocean sound monitoring from state-of-the-art observation platforms with constrained resources. Hydrophones, as well as the vast majority of marine sensors, do not have standardized interfaces, proving difficult to integrate into observation platforms. To overcome this lack interoperability, users develop their specific drivers, which is a time-consuming and error-prone task that requires in-depth knowledge of both the sensor and the observation platform. Furthermore, since environmental data should be shared across the scientific community to be properly analyzed, data has to be post-processed and formatted in order to be published in spatial data infrastructures. If a standardized approach is not considered, these operations require custom-made software components which increase the operation and maintenance costs of both acquisition systems and data infrastructures. This thesis applies interoperability techniques to reduce the operational costs of ocean sound monitoring. First, a standardsbased interoperable architecture to integrate hydrophones (and other marine sensors) into observation platforms is proposed. Standardized data and metadata management is carefully considered within this architecture, ensuring its reusability and easing its coupling into spatial data infrastructures. An efficient, cross-platform algorithm for in situ ocean sound processing is provided. This algorithm allows state-of-the-art observation platforms to monitor ocean sound in realtime, regardless of the platform's constraints. Sempre hi ha hagut so de fons en els oceans degut a factors naturals. Tanmateix, des dels primers desplegaments d'hidròfons a mitjans del segle XX, els nivells de so mesurats als oceans han anat augmentant contínuament degut a l'increment d'activitats humanes com la navegació, la construcció, els sonars i les campanyes sísmiques d’exploració. El so en els oceans és un important factor ambiental per moltes espècies, especialment per aquelles utilitzant-lo per localització i comunicacions (per exemple mamífers marins). Per tant, la introducció d'energia a l'habitat marí ha de ser degudament mesurada. La intrínseca naturalesa de l’entorn marí complica l’adquisició de conjunts de dades representatius sobre el so als oceans. Grans extensions d’aigua a diferents profunditats han de ser mesurades, limitant considerablement els mètodes de mostreig disponibles. El desplegament d’estacions fixes, com gravadors autònoms o conjunts d’hidròfons fondejats són cars i presenten problemes com la limitada autonomia o el reduït ample de banda disponible. Els vehicles no tripulats d’última generació, com els planadors submarins o vehicles de superfície autònoms, tenen una autonomia excepcional, al mateix temps que poden cobrir extenses àrees. Tanmateix, aquestes plataformes d’observació encara tenen més limitacions en termes d’energia i comunicacions. Els hidròfons comercials utilitzats en el control del so als oceans, proporcionen dades acústiques en brut en gravacions o bé en transmissió contínua. Per tant, aquestes dades en brut han de ser processades a les estacions de terra. Degut al gran volum de dades produït pels hidròfons, la transmissió de les dades en brut utilitzant comunicacions amb ample de banda reduït no és possible. Malgrat això, si les dades es poguessin processar a bord de les plataformes d’observació, la quantitat d’informació a ser transmesa es reduiria en varis ordres de magnitud. Els hidròfons, així com la gran majoria de sensors marins, no tenen interfícies de comunicacions estàndards, dificultant la seva integració en plataformes d’observació. Per a superar aquesta falta d’interoperabilitat, els usuaris des envolupen programes específics, resultant en una tasca propensa als errors, lenta i laboriosa. Desenvolupar aquests programes requereix coneixement en profunditat tant de la plataforma d’observació com del sensor. A més a més, les dades ambientals han de ser compartides entre la comunitat científica per a ser degudament analitzades. Per tant, aquestes dades han de ser post-processades i se’ls hi ha de donar un format adequat per a ser publicades en infraestructures de dades científiques. Si no s’utilitza una estratègia estandarditzada, aquestes operacions requereixen components fets a mida, augmentant significativament els cos tos d’operació i manteniment dels sistemes d’adquisició i les infraestructures de dades. Aquesta tesis aplica tècniques d’interoperabilitat per a reduir els costos operacionals dels sistemes d’observació de so oceànic. Primer, es proposa una arquitectura universal, estandarditzada i interoperable per la integració d’hidròfons (i altres sensors marins ) en plataformes d’observació marines. Dins d’aquesta arquitectura es posa èmfasis en la gestió de dades i metadades de manera estàndard, assegurant la reutilització, modularitat i capacitat d’acoblament a infraestructures de dades. Respecte al so submarí, es proposa un algoritme multiplataforma per al processat a bord del so oceànic. Aquest algoritme permet a les plataformes d’observació d'última generació realitzar mesures en temps real de so oceànic, superant les seves limitacions intrínseques.

show abstract

The EGIM, a generic instrumental module to equip EMSO observatories

Cited by 6 publications

References 1 publication

Deep learning based deep-sea automatic image enhancement and animal species classification

Deep learning based deep-sea automatic image enhancement and animal species classification

Metadata-Driven Universal Real-Time Ocean Sound Measurement Architecture

Marine sensor interoperability applied to in situ ocean sound processing

Contact Info

Product

Resources

About