An underwater acoustic target recognition method based on auditory perception feature deep learning

Abstract

The invention provides an underwater acoustic target recognition method based on auditory perception feature deep learning, which aims at the current situations of low signal-to-noise ratio of underwater acoustic data, difficulty in separable feature extraction, poor recognition generalization ability and the like; and the method comprises the following steps: firstly, carrying out preprocessing based on an auditory perception method to generate an MFCC spectrogram sample and a GFCC spectrogram sample; for said spectrogram samples, constructing a deep network model based on a deep learning method for processing; finally carrying out weighted joint judgment on a multi-model result, and outputting a target recognition result. According to the method, the advantages of an auditory perception method in feature representation and a deep learning method in feature abstraction are comprehensively utilized, and advantage complementation among multi-dimensional auditory perception features is realized through a multi-model joint judgment method, so that the target recognition robustness is improved; the invention is an innovative method for application of an artificial intelligence algorithm in the field of underwater acoustic signal processing.

Description

technical field [0001] The invention belongs to the technical field of underwater target recognition and artificial intelligence, and mainly relates to an underwater acoustic target recognition method based on deep learning of auditory perception features. Background technique [0002] The identification of radiation noise of underwater targets is one of the main functions of sonar, which can provide an important basis for comprehensive decision-making of sonar operators. Affected by factors such as complex target noise generation mechanism, time-space variable transmission of ocean channels, strong multi-target interference, platform background noise, and difficulty in obtaining high-quality data, underwater acoustic target noise recognition has long been an internationally recognized problem that needs to be solved urgently. [0003] After long-term evolution, the human auditory system has a strong ability to analyze and identify sound signals. By referring to the successf...

AI Technical Summary

Problems solved by technology

For underwater acoustic target recognition, many research teams at home and abroad have carried out application research on deep learning methods, but generally the models used are relatively single, and no in-depth research has been conducted on the characteristics of underwater acoustic target recognition.

Images