A vision-based system for on-board identification and estimation of discarded bio-mass: A tool for contributing to marine resources sustainability

摘要

In marine harvesting, fish waste due to the discards of non-targeted species represent a risk for the sustainability of fisheries, a loss of potentially valuable living resources and a threat for the ecological equilibrium. The project FAROS aims to develop and implement an efficient, integral management network for discards, to reduce global impact of this established practice, while putting in value and optimising the use of the by-catch. The development of an on-ship automated system for discards evaluation and species biomass estimation is proposed in this contribution. The Biomass Estimation Optical System (BEOS) integrates machine vision technologies and optical information processing and feature extraction by means of nonlinear modeling based on artificial neural networks. A first functional prototype and a methodology for system implementation and operation are presented here.