Project: Automated sea lice detection by computer vision.

Sea lice (Lepeophtheirus salmonis) are a very significant problem in salmon aquaculture. They cost the_x000D_Scottish and Norwegian salmon farming industry around €75m per year, compromise fish welfare, can cause_x000D_pathological damage to salmon and are perceived as an environmental threat. Increasingly effective control_x000D_of lice is vital for the salmon industry. Failure to achieve this will have a high financial and political_x000D_price which could threaten the industry’s long term interests. The goal of this project is to develop a tool that_x000D_will assist the industry in the control of lice through improved knowledge of the lice burdon. If it achieves this it will directly benefit fish welfare, the environment and the fish farming industry._x000D__x000D_Effective and environmentally responsible control of sea lice depends on an accurate estimate of the lice population on a farm. This is currently achieved only by close manual inspection of anaesthetised fish. This approach is labour intensive and very dependant on the skill of the inspector. It is of limited statistical validity because of the small sample sizes possible. The crowding of fish needed to collect the fish sample can also compromise fish welfare and limit the frequency of inspection._x000D__x000D_In this project we will develop an alternative monitoring method which will automatically detect and count lice on fish swimming freely in the sea pens. This is a camera based approach using the natural fluorescent of sea lice to enhance their visibility. Images of the fish are automatically collected as they swim through a monitoring station and the colour information in these images is inspected for the unique signature of fluorescing lice. Image analysis software is used to segment the images, count the lice and identify their location on the fish._x000D__x000D_The feasibility of this approach was investigated in a preliminary study (Link feasibility study LK0663F: Passive monitoring of sea lice) supported by the Scottish Office. It was shown to be challenging but probably achievable goal. Marine Harvest, the worlds largest salmon farming company contributed to this feasibility investigation and will continue to support this development by providing access to fish, sea lice expertise and practical assistance during experimental trials. _x000D__x000D_The project aims to solve the reCOing technical problems identified by the feasibility study and to develop prototype hardware to the point where the lead company can take it through to a commercial product. It is envisaged that this product will comprise one or more portable imaging frames linked to a central computer. These imaging frames will be suspended at various depths on fish farms so that the fish can swim through them. As this happens, the fish will be detected and images will be collected automatically. This data will be passed to the central computer which over a period of hours or days will build up an estimate of the lice population on which intervention decisions can be based. _x000D__x000D_If the aims of this project are achieved it will benefit the commercial salmon industry, the welfare of the fish and the environment. The salmon farmers associations and large scale operators in the industry have confirmed their interest in the project._x000D__x000D_The project consortium comprises Vaki, Silsoe Livestock Systems, Marine Harvest (Scotland) and the University of Strathclyde. _x000D__x000D_Vaki Ltd. is a well established and leading supplier of monitoring equipment to the fish farming industry._x000D__x000D_Silsoe Livestock Systems is a research company which focuses on novel engineering developments for improving the monitoring and welfare of farmed fish and livestock. _x000D__x000D_Marine Harvest is the worlds largest salmon farming company. Marine Harvest Ltd. does not require funding but will contribute to this project by providing access to fish and farm sites during the two phases of equipment testing, providing also local support, on-site management and lice counts as needed. _x000D__x000D_The University of Strathclyde has an international reputation for its research in lice population modelling. Strathclyde research will add value to the project by providing research identifying the best ways to use the data provided by this automatic lice counting system, and also establishing how these results are best compared with the data currently obtained by manual inspection. The Strathclyde contribution comes at the end of the development when the level of detail in the data being generated is known. Strathclyde will seek funding from an external funding source for this work . Potential funders include the UK research council (BBSRC) and Norwegian Fishery and Aquaculture Research Fund (FHF). Both of these funders have already expressed interest in this development. _x000D__x000D_

Acronym	VisuaLice (Reference Number: 4721)
Duration	01/06/2009 - 31/12/2012
Project Topic	Accurate lice counts on salmon farms are vital for environmentally responsible control. This is currently done by manual inspection of anaesthetised fish. We will develop an automatic, remote system to inspect fish as they swim freely in the pen, so improving fish welfare and benefiting the industry
Project Results (after finalisation)	UPEI supplied expertise to this the project on the basis of funding provided by a third party. It did not receive direct Eurostars funding support and does not expect to benefit further from the project. _x000D__x000D_In this project a system to visualise lice on farmed fish was created.
Network	Eurostars
Call	Eurostars Cut-Off 2

Project partner

Number	Name	Role	Country
5	Marine Harvest (Scotland) Ltd	Partner	United Kingdom
5	Silsoe Livestock Systems Ltd	Partner	United Kingdom
5	University of Prince Edward Island, Department of Health Management, Atlantic Veterinary College	Partner	Canada
5	University of Strathclyde	Observer	United Kingdom
5	Vaki Aquacultural Systems Ltd.	Coordinator	Iceland