Practical implementation of real-time fish classification from acoustic broadband echo sounder data- RealFishEcho progress report: Year 1-June 2017

The EU has by means of new policy restricted the discarding of fish at sea, including bycatch and slipping. As a result, the fishing industry now requires improved methods to identify fish species and size before the catch process begins. With the introduction of robust broadband echosounder (i.e. downward looking sonar) on the market, active acoustic data now have the potential for improved target discrimination for both type and size (i.e. determine fish species and size). The aim of this project is to develop methods for fish classification and size estimation using data from broadband echosounder and further implement them into a software for near-real time viewing. Such a software would help skippers to take better informed decisions while fishing. This project is running over three years (June 2016 to June 2019) in collaboration between Wageningen Marine Research (WMR), Redersvereniging voor de Zeevisserij (RVZ) and TNO. This document reports on the improvement of the methods used for fish species classification