Artificial neural networks as a tool for species identification of fish schools

Abstract

Fish schools of sardine, anchovy, and horse mackerel can be discriminated from each other, under given conditions, using a set of parameters extracted from echo-integration data. Trawl sampling and hydroacoustic data were collected in 1992 and 1993 in the Thermaikos Gulf by using a towed dual-beam 120 kHz transducer. The parameters extracted from the available schools were used to train multi-layered feed-forward artificial neural networks. Various applied networks easily generated associations between school descriptors and species identity, providing a powerful tool for classification. The expertise of the trained network was tested with data from identified schools not used in training. The use of neural networks cannot replace classical statistical procedures, but offers an alternative when there are significant overlaps in the school characteristics and the parametric assumptions are not satisfied.