On statistical linear inverse problems in fishery acoustics

The paper summarizes several aspects of solving statistical linear inverse problems that occur in the processes related to fish abundance estimation performed on the basis of acoustically acquired data. As the difficulty in solving such problems inherently comes from uncertainty in determination of so called inversion kernel, the paper discusses the key approaches that lead to formulation of statistical relations between acoustic echoes observed from swimming fish in their natural environment and mostly required fish parameter, namely physical length of the fish.