Prosjektnummer
900594
Extension of the Eurostar’s VisuaLice E!4721 project to evaluate and exploit automatically collected lice data / Delprosjekt til VisuaLice E!4721: Statistisk evaluering av ny tellemetode for lakselus
Prosjektet ble etablert som et tillegg til et prosjekt finansiert av næringsaktører og EU-programmet EuroStars, der målet var å utvikle en teknologi for automatisk overvåking og telling av lakselus på laks i merder. Dette prosjektet skulle bidra med data fra automatiske tellinger i felt til forskning finansiert av FHF ved University of Prince Edward Island i Canada, for utvikling av en statistisk modell for tolking og utnyttelse av dataene.
Siden det så langt ikke har lyktes å nå frem med utviklingen av den aktuelle teknologien for telling, ble FHF-prosjektet avsluttet våren 2013. Det foreligger likevel en teoretisk modell som simulerer effekten av ulike strategier for prøvetaking basert på manuell og automatisk overvåking. Modellen inkluderer en lusepopulasjon, og en forenklet fire-stadiers livssyklus (egg - copepoditter - juvenile - voksne inkludert gravide hunnlus). En tilleggsparameter modellerer introduksjon av copepoditter fra eksterne kilder (f.eks. andre anlegg eller villfisk). Lusemengden på hver fisk kan følges gjennom en toårig produksjonssyklus.
Siden prosjektet ikke fikk relevante data fra automatisk registrering i felt er ikke modellen ferdig testet, men videre utvikling og verifisering vil kunne skje i samarbeid med bedrifter som utvikler automatiske tellesystemer for lakselus.
Siden det så langt ikke har lyktes å nå frem med utviklingen av den aktuelle teknologien for telling, ble FHF-prosjektet avsluttet våren 2013. Det foreligger likevel en teoretisk modell som simulerer effekten av ulike strategier for prøvetaking basert på manuell og automatisk overvåking. Modellen inkluderer en lusepopulasjon, og en forenklet fire-stadiers livssyklus (egg - copepoditter - juvenile - voksne inkludert gravide hunnlus). En tilleggsparameter modellerer introduksjon av copepoditter fra eksterne kilder (f.eks. andre anlegg eller villfisk). Lusemengden på hver fisk kan følges gjennom en toårig produksjonssyklus.
Siden prosjektet ikke fikk relevante data fra automatisk registrering i felt er ikke modellen ferdig testet, men videre utvikling og verifisering vil kunne skje i samarbeid med bedrifter som utvikler automatiske tellesystemer for lakselus.
-
Report: VisuaLice: Population Interpretation of Passive Sea Lice Monitoring
University of Prince Edward Island. March 2013. By Ruth Cox and Crawford W. Revie.
Background
Vaki Aquaculture Systems (Iceland), Marine Harvest (Scotland/Norway), Silsoe Livestock Systems (UK) and the University of Prince Edwards Island (Canada) are collaborating in Eurostars project “VisuaLice E!4721” on the development of equipment to automatically detect and count sea lice. The data generated by this equipment will be similar to, but not directly comparable with, that generated by manual counting. It will however have more statistical reliability, be more repeatable and has the potential to provide levels of detail not previously available on sea lice population dynamics which will facilitate more effective lice control.
Vaki Aquaculture Systems (Iceland), Marine Harvest (Scotland/Norway), Silsoe Livestock Systems (UK) and the University of Prince Edwards Island (Canada) are collaborating in Eurostars project “VisuaLice E!4721” on the development of equipment to automatically detect and count sea lice. The data generated by this equipment will be similar to, but not directly comparable with, that generated by manual counting. It will however have more statistical reliability, be more repeatable and has the potential to provide levels of detail not previously available on sea lice population dynamics which will facilitate more effective lice control.
Objectives
To evaluate and exploit automatically collected lice data / Å dokumentere telleresultater oppnådd med automatisert telling og med dagens manuelle tellemetoder.
To evaluate and exploit automatically collected lice data / Å dokumentere telleresultater oppnådd med automatisert telling og med dagens manuelle tellemetoder.
Expected project impact
Regular sea lice counting is a vital part of lice control. It is however a manual process which is time consuming and very dependent for its accuracy on the skill of the inspector and access to the sea pens.
Crowding the fish to select the sample imposes additional stress on the fish. Because of the time required only a small number of fish can be sampled, however as lice numbers are driven down increasingly large samples will be required to maintain the statistical reliability of the population estimates. A passive, automated counting system offers the benefits of enhanced repeatability and accuracy, larger sample sizes, continuous monitoring, lower costs and lower levels of disturbance for the fish.
In Norwegian
Vitenskapelig dokumentasjon av telleresultater oppnådd med automatisert telling og dagens manuelle telling vil, hvis de blir tilfredstillende, bidra til at automatisert telling av lakselus kan bli en realitet i havbruksnæringen innen kort tid.
Regular sea lice counting is a vital part of lice control. It is however a manual process which is time consuming and very dependent for its accuracy on the skill of the inspector and access to the sea pens.
Crowding the fish to select the sample imposes additional stress on the fish. Because of the time required only a small number of fish can be sampled, however as lice numbers are driven down increasingly large samples will be required to maintain the statistical reliability of the population estimates. A passive, automated counting system offers the benefits of enhanced repeatability and accuracy, larger sample sizes, continuous monitoring, lower costs and lower levels of disturbance for the fish.
In Norwegian
Vitenskapelig dokumentasjon av telleresultater oppnådd med automatisert telling og dagens manuelle telling vil, hvis de blir tilfredstillende, bidra til at automatisert telling av lakselus kan bli en realitet i havbruksnæringen innen kort tid.
Project design and implementation
An approach to automatic lice counting was investigated in 2005–2007 in a project led by Silsoe in collaboration with Marine Harvest and University of Strathclyde. The fluorescence characteristics of lice were investigated and colour signatures in the collected images which could be used to detect lice were identified. This equipment has been further developed, and prototype equipment will be tested in commercial situations on farms in both Norway and Scotland early in 2011. The main objective of the testing will be to broaden the range of environments it is able to function in and to collect comparative data from the automatic monitoring system as well as frequent manual lice counts for use by UPEI in the data validation and
modelling research.
An approach to automatic lice counting was investigated in 2005–2007 in a project led by Silsoe in collaboration with Marine Harvest and University of Strathclyde. The fluorescence characteristics of lice were investigated and colour signatures in the collected images which could be used to detect lice were identified. This equipment has been further developed, and prototype equipment will be tested in commercial situations on farms in both Norway and Scotland early in 2011. The main objective of the testing will be to broaden the range of environments it is able to function in and to collect comparative data from the automatic monitoring system as well as frequent manual lice counts for use by UPEI in the data validation and
modelling research.
In this sub-project the research will explore how the novel data sets generated can be best exploited. Statistical population modelling using real world scenarios based on historical lice data from Norwegian and Scottish farms, as well as data provided by the equipment trials, will be used to examine how valid comparisons can be made between lice infestation data generated by existing manual inspection and data generated by the new technology. New features of this data should facilitate better estimates of rates of population change, short term population variations and enhanced accuracy of prevalence estimation which could facilitate a move from abundance modelling to prevalence modelling. Researchers modelling sea lice population dynamics have found it difficult to accurately parameterise lice growth and, subsequent to treatment, decay curves. The increased accuracy and temporal resolution provided by this novel approach will offer many benefits to this type of modelling work.
Dissemination of project results
A peer‐reviewed scientific publication summarizing issues around validity, comparability and utility of this new approach to surveillance will be a key outcome of this element of the research programme. In addition it is proposed that a sea lice sampling ‘toolkit’ which will be constructed to validate the new technique will be placed in the public domain. This toolkit will allow users to simulate different infestation scenarios and will be of more general value in that a range of alternative surveillance scenarios can be explored.
A peer‐reviewed scientific publication summarizing issues around validity, comparability and utility of this new approach to surveillance will be a key outcome of this element of the research programme. In addition it is proposed that a sea lice sampling ‘toolkit’ which will be constructed to validate the new technique will be placed in the public domain. This toolkit will allow users to simulate different infestation scenarios and will be of more general value in that a range of alternative surveillance scenarios can be explored.