Oral Presentation Australian Society for Fish Biology Conference 2024

Machine learningĀ applications to electronic monitoring in Australia's marine fisheries (111482)

Geoff Tuck 1 , Richard Little 1 , Dadong Wang 2
  1. CSIRO, Sustainable Marine Futures, Hobart, Tasmania, Australia
  2. CSIRO, Data 61, Marsfield, NSW, Tasmania

Fisheries worldwide are shifting towards on-vessel cameras, electronic monitoring (EM), for catch monitoring, with machine learning and artificial intelligence (MLAI) applications expected to follow. The impact of on-vessel species identification will be significant. The high costs of human observers and EM reviewing for both government and industry highlight the need for automated species identification using MLAI, which can reduce costs and increase monitoring coverage across entire fleets. This will lead to improved governance, management, and environmental outcomes. The CSIRO Marine Visual Technologies (MVT) team has been actively developing MLAI technology for fishery management, including algorithms for event detection and species identification of target and non-target species. The workflow also includes tracking, image enhancement, and a cloud-based reporting and auditing interface.