The University of Los Lagos will develop an underwater computer vision and Machine Learning system capable of automatically detecting the smoltification status of Coho salmon in productive conditions. The initiative continues a FONDEF IDeA research developed by the University of Los Lagos since 2023.
The University of Los Lagos won the FONDEF IT26I0063 project SmoltSense: An underwater system based on computer vision for the automatic detection of the smoltification status in Coho salmon, an initiative that seeks to incorporate artificial intelligence and computer vision into the monitoring of one of the relevant processes for salmon farming.
The project is led by Óscar Mora Garay, belonging to the Aquaculture Production Unit (APU) from the University of Los Lagos, and includes the participation of Southern Smart Engineering Solutions SpA (3Se) as an associated entity. The University of Los Lagos participates as the main beneficiary.
Beyond the surface: technology and artificial intelligence to observe fish
The proposal considers the development of an integrated underwater capture and image analysis system, using immersion cameras, controlled lighting and a transmission and power architecture designed to operate in farming units.
From the images obtained underwater, machine learning algorithms will allow automatically classifying fish according to their smolt or non-smolt condition, moving from individual assessment to population monitoring. The project also includes the development of a software platform for the registration, storage and visualization of the information generated by the system, including the number of fish sampled and the percentage estimate of the mollt state within the cultivation units.
In this way, the project seeks to transform the images captured in the aquatic environment into information that can be used to support productive decisions.
Continuity of an investigation conducted in ULagos
The new initiative represents the continuation of the project FONDEF IDeA 2023 ID23I10131, called “Development of a productive tool based on Machine Learning for the determination of the state of smoltification of Atlantic salmon using real-time image analysis.”
This new phase will make it possible to move from the development of an image analysis tool to an integrated underwater system, oriented to its evaluation under operational conditions. The project envisages a stage of evaluation of the performance and applicability of the technological solution, considering its accuracy and robustness as a tool to support the making of productive decisions associated with the transfer of the fish to the sea.
Artificial intelligence for aquaculture
The initiative is developed in the priority area of Fisheries and Aquaculture and is linked to the OECD discipline of Information Science and Bioinformatics, articulating research and technological development capacities around a specific challenge of the aquaculture industry.
For the University of Los Lagos, the award strengthens an applied research line that connects aquaculture, artificial intelligence, computer vision and technology transfer, generating capacities for the development of solutions oriented to the needs of the territory and the productive sector.
The University of Los Lagos continues to advance in the development of technologies that allow for the incorporation of new automated analysis tools into aquaculture, seeking that information obtained directly from the cultivation units can be transformed into useful knowledge for more timely and data-based productive decision-making.
