AutoFish: Automatic Fish Phenotyping Tool for Sustainable Aquaculture and Smart Fisheries (Old ID 28947)
Role
Principal Investigator
Description
This project aims to advance an initial implementation of a new automatic fish phenotyping tool, named AutoFish, that will enhance aquaculture farming and post-harvest processing practices by accelerating the collection of data and automatically analysing and leveraging it using the latest advances in computer vision and machine learning technologies. We have already developed a ground-breaking solution for the aquaculture industry, and produced a Proof Of Concept (POC) device, which has been successfully tested. This POC is ready to get tailored, integrated and/or retrofitted into an available fish and aquaculture farm for grading, phenotypic collection and/or processing lines, due to its modular stand-alone nature. The AutoFish concept has been proven with Barramundi as a test species. However, it can be used to collect images of, and trained to predict features of other fish and/or aquaculture species. AutoFish can also be trained to detect and predict any customertailored features and traits of the fish/product, and/or be used to assess the health of fish/product, if th health issues are visually recognisable by an RGB camera and the background phenotypic data present for training. Hence this technology would be scalable and can be implemented for other aquaculture products worldwide. The main aim of this project is to unlock the potential of our initial AutoFish POC by applying it to, or altering it for, a partner Barramundi (or Prawn) farm, to solve a real end-users’ problem. This problem can be about any monitoring aspect, as long as cameras can capture it. We have narrowed down the scope of this project to Barramundi (and/or Prawn) due to our wider industry network in these two species.
Date
01 Mar 2023 - 30 Jun 2023
Project Type
GRANT
Keywords
Fish morphology;Deep Learning;Fish monitoring;Machine Vision
Funding Body
James Cook University
Amount
50000
Project Team
Alzayat Saleh;Dean Jerry;Dave Jones