Imaging sonar tracking of salmon for size and tail beat frequency

Abstract

In order to set sustainable harvest limits on anadro-mous fish, fishery managers require a method for estimating return numbers of individual species. As seasonal migration times of several returning species may overlap, simple fish counts are insufficient. This paper presents an algorithm for tracking fish crossing an insonified area using an extended Gaussian mixture probability hypothesis density filter as well as analyzing both their lengths and the frequency of their tail beats. The combination of the two provides a useful estimate of the return size of a given fish species when combined with seasonal return timing. The algorithm is tested against data collected using a dual frequency imaging sonar on the Kenai River in Alaska and compares favorably to estimates made by manual inspection and sub-sampling.