A general procedure for estimating the composition of fish school clustersusing standard acoustic survey data

An algorithm to identify classes of fish in acoustic backscatter images wou ld improve the accuracy of acoustic biomass estimates over manually scrutin ized images. A generalized Bayesian procedure for such identification calle d BASCET is presented, and two implementation strategies for the procedure are compared using simulated acoustic survey data. The procedure has severa l unusual characteristics: it evaluates schools not individually but in clu sters; it makes use of human experience at cluster identification; it prese nts measures of uncertainty in all estimation results; and it constructs th e training set required for supervised learning automatically using spatial and temporal assumptions. The simulation study comparison suggests that ma king use of temporal and spatial structure in the acoustic data leads to im proved estimation performance. On the simulated data, the BASCET algorithm correctly identified the dominant fish class in 15 of 16 cases. However, th e simulation model generates acoustic survey data based on the same assumpt ions used in BASCET, assumptions that may differ from a real acoustic surve y. The study also assumed that the human experience incorporated in the Bay esian prior distributions was not misleading. Performance of BASCET on real acoustic data is presented in a companion paper.