Automatic classification of acoustically detected krill aggregations: A case study from Southern Ocean

Acoustic surveys represent the standard methodology to assess spatial distribution and abundance of pelagic organisms characterized by aggregative behaviour. The species identification acoustically observed aggregations is usually performed taking into account biological sampling according expert-based knowledge. precision survey estimates, such as total distribution, strongly depends on e?ciency acoustic well identi?cation. In this context, automatic speci?c groups based energetic morphological features could improve identi?cation process, allowing estimates or overcome problems related biases in sampling. present study, we test use well-known unsupervised clustering methods focusing two important krill namely Euphausia superba crystallorophias . order obtain a reference classi?cation, echoes were ?rst classi?ed criteria parameters accounting for response at 38 kHz 120 kHz. Di?erent combined with three distance metrics then tested working wider set parameters, depth insonified aggregation features. performances evaluated comparing classi?cation one obtained clustering. Obtained results showed that k-means performs better than considered hierarchical methods. Our ?ndings also evidenced variables rather all available ones highly impact performances. • Species an step provide reliable information about marine organisms. looking Krill can be classified considering specific Different algorithms used identify detected species. K-means classification acceptable accordance regression same purpose.