The ‘Automated Fish Identification’ Project
Harrier Project Management; SeaGIS; In-situ Marine Optics; The Business Research and Innovation Initiative
Australian fisheries are worth $2.5 billion per year, forming an important food source and sociocultural resource. In order to safeguard this resource, detailed and ongoing monitoring of fish stocks is required. However, the manual counting of fish involved in monitoring programs is costly and time-consuming.
By combining ‘Baited Remote Underwater Video Systems’ and machine-learning algorithms, the process of fish counting can be automated to make it both more cost and time efficient.
The CIC, with industry support, developed and tested an automated video-based fish counting solution.
Figure 1: An image of the system at work.