Covariance Mean Variance Classification (CMVC) techniques: Application to the acoustic classification of zooplankton

Accurate acoustic classification of zooplankton species has the potential to significantly improve estimates of zooplankton biomass made from ocean acoustic backscatter measurements. Theoretical models have been developed for three zooplankton scattering classes (hard elastic shelled, e.g., pteropods; fluidlike, e.g., euphausiids; gas bearing, e.g., siphonophores), providing a sound basis for a model-based classification approach. The Covariance Mean Variance Classification (CMVC) techniques classify broadband echoes from individual zooplankton based on comparisons of observed echo spectra to model space realizations. Three different CMVC algorithms were developed: the Integrated Score Classifier, the Pairwise Score Classifier, and the Bayesian Probability Classifier; these classifiers assign observations to a class based on similarities in covariance, mean, and variance, while accounting for model space ambiguity and validity. The CMVC techniques were applied to several hundred broadband (∼350–750 kHz) echoes acquired from 24 different zooplankton to invert for scatterer class. All three classification algorithms had a high rate of success with high-quality, low SNR data. The CMVC approach was also applied to several thousand echoes from fluidlike zooplankton to invert for angle of orientation using both theoretical and empirical model spaces; excellent success rates were achieved with the empirical model spaces.