Aquaculture Europe 2021

October 4 - 7, 2021

Funchal, Madeira

Add To Calendar 07/10/2021 11:20:0007/10/2021 11:40:00Europe/LisbonAquaculture Europe 2021HOW MACHINE LEARNING AND ARTIFICAL INTELLIGENCE HAVE SIGNIFICANTLY IMPROVED OPEN OCEAN FISH FARMINGBerlim-HotelThe European Aquaculture Societywebmaster@aquaeas.orgfalseDD/MM/YYYYaaVZHLXMfzTRLzDrHmAi181982

HOW MACHINE LEARNING AND ARTIFICAL INTELLIGENCE HAVE SIGNIFICANTLY IMPROVED OPEN OCEAN FISH FARMING

 

Langley Gace *

Innovasea Systems, Inc.

425 Ericksen Avenue NE

Suite 101

Bainbridge Island, WA 98110

LGace@InnovaSea.com



Open ocean fish farming has changed dramatically in the past few years.  The industry is embracing the significant advantages that machine learning and artificial intelligence (AI) have to offer as well as affordable high definition underwater cameras.  Feeding submerged pens is one operation that has benefited significantly from these technological advancements.

Efficient feeding is critical to a farm’s profitability as well as reducing its impact on the local environment.  Feed- related expenses are often responsible for 50% or more of a farm’s operating costs.  The 10% improvement in the FCRe (2.0-1.8) can potentially equal more than $1.1 million savings annually on a 3000 ton farm.

Until recently,  all  feed was delivered via  forced air from the feed barge to the pen located at the surface.  With the advent of submerged pens, feed then had to be delivered via a slurry of water.  In addition to requiring less power, there are several other advantages, such as providing a cushion of fluid to deliver more of the pellet to the pen.

High-definition underwater cameras are no longer cost-prohibitive, allowing the farmer to observe feeding and other activities in all the pens from the safety and comfort of their office on the feed barge or remotely on land.  Video from these cameras are analyzed by AI algorithms to identify pellet waste and fish satiation.  These programs help the feed manager to reduce feed waste, optimize feed rate, and to determine when the fish have been adequately fed.

Stereoscopic biomass estimation cameras constantly measure fish dimensions within the pen.  The onboard algorithms then provide any near real- time biomass estimation within 3% accuracy.  This data is critical to the farm managers enabling them  to optimize feed schedules as well as predict harvest mass and schedules.