Whale and dolphin Click-like sound and traditional sonar signal classification method and device

Abstract

The invention belongs to the field of marine applied acoustics. In order to provide a new artificial classification method for underwater acoustical signals with stable classification results, low workload, good subjectivity, accurate feature extraction, high classification accuracy, and good adaptability, the invention provides a whale and dolphin Click-like sound and traditional sonar signal classification method and device. The device includes three subsystems: a signal preprocessing subsystem, a signal time-frequency feature extraction subsystem, and a signal time-frequency feature classification subsystem. The signal preprocessing subsystem comprises: a hydrophone, a sound storage medium, a sound de-noising module and an endpoint detection module. The signal time-frequency feature extraction subsystem comprises a signal short-time Fourier transform preprocessing module, a signal time-frequency contour extraction module, a signal time-frequency contour curve fitting and signal time-frequency feature extraction module. The whale and dolphin Click-like sound and traditional sonar signal classification method and device are mainly applied to marine applied acoustics.

Description

technical field [0001] The invention belongs to the field of marine applied acoustics, in particular to a method and a device for classifying cetacean click-like call pulses and traditional sonar signals. Background technique [0002] In marine passive acoustic monitoring, underwater sonar detection or underwater acoustic communication, cetacean click-like sound pulse signals and traditional sonar signals are very likely to appear in the same sea area at the same time, and are received and received by the acoustic monitoring network or sonar receiving equipment together. Record. Among them, traditional sonar signals mainly include Continuous Wave (CW for short) signal, Linear Frequency Modulation (LFM for short) signal and Hyperbolic Frequency Modulation (HFM for short) signal. Effectively identify and classify cetacean click-like call pulse signals and traditional sonar signals (CW signals, LFM signals and HFM signals) in the received and recorded sounds, which is helpful ...

AI Technical Summary

Problems solved by technology

The current underwater acoustic signal recognition and classification still mainly adopts the method of manual classification, and the signal classification is realized by listening and observing the signal waveform and the signal spectrum characteristics with the help of analysis tools. This method is highly subjective and the classification accuracy fluctuates greatly.

In addition, with the increase in the scale of the passive acoustic monitoring network, the formation of artificial sonar arrays and the complexity of the underwater communication structure, the amount of data to be processed continues to increase, and it will be difficult to ensure the real-time and accuracy of the data using manual classification methods

However, the existing automatic classification methods for underwater sound signals are difficult to accurately extract signal features, the classification accuracy is low, and it is difficult to adapt to the complex and changeable marine environment.

Images