Spectral features of hydroacoustic noises in determination of class of underwater objects

Abstract

Decision-making systems based on the analysis of big data and machine learning methods are successfully applied in many areas. It is necessary for correct classification of an object (or action) to determine its features helping to achieve the necessary classification accuracy. The machine learning method was approved analytically and experimentally   for solution of the problem of classification of underwater objects using spectral features of hydroacoustic noises. The training of the model with a bid acoustic data and the development of a sensor network will allow to create a system of quick response to the unauthorized access (boats and persons) to protected water area.