Classification Of Underwater Acoustic Signals Research Articles

Underwater acoustic signal classification based on a spatial–temporal fusion neural network

Underwater acoustic signal classification based on a spatial–temporal fusion neural network

Self-Supervised Learning-For Underwater Acoustic Signal Classification With Mixup

Self-Supervised Learning-For Underwater Acoustic Signal Classification With Mixup

Self-supervised learning–based underwater acoustical signal classification via mask modeling

The classification of underwater acoustic signals has garnered a great deal of attention in recent years due to its potential applications in military and civilian contexts. While deep neural networks have emerged as the preferred method for this task, the representation of the signals plays a crucial role in determining the performance of the classification. However, the representation of underwater acoustic signals remains an under-explored area. In addition, the annotation of large-scale datasets for the training of deep networks is a challenging and expensive task. To tackle these challenges, we propose a novel self-supervised representation learning method for underwater acoustic signal classification. Our approach consists of two stages: a pretext learning stage using unlabeled data and a downstream fine-tuning stage using a small amount of labeled data. The pretext learning stage involves randomly masking the log Mel spectrogram and reconstructing the masked part using the Swin Transformer architecture. This allows us to learn a general representation of the acoustic signal. Our method achieves a classification accuracy of 80.22% on the DeepShip dataset, outperforming or matching previous competitive methods. Furthermore, our classification method demonstrates good performance in low signal-to-noise ratio or few-shot settings.

MSLEFC: A low-frequency focused underwater acoustic signal classification and analysis system

Many feature extraction techniques are employed to study the classification of underwater acoustic, however, most of these techniques result in the loss of detailed information in this field. A Multi-scale short-time Fourier transform (MS-STFT) method was proposed to improve the low-frequency information and maintain the detailed information by increasing the number of channels. An effective data augmentation approach was proposed and a Ladder-like Encode (LE) architecture was built to increase the generalizability of the model and the classification accuracy for the features extracted. Finally, A Frequency-CAM (FC) method is proposed to analyze the frequency band locations that the neural network is interested in when conducting classification tasks on various classes of data. The integration of the above technologies is called MSLEFC. The performance of this system has been tested on two ship radiation noise datasets and obtained 82.94% and 96.06% accuracy, respectively. On the ShipsEar dataset, the proposed method increases the accuracy by 0.7% over the previous state-of-the-art method. The proposed model architecture has significant improvements in parameter size and computation compared to ResNet50.

A Speech recognition of isolated words in the “Baoulé” language via back-propagation neural networks (BNN)

The main objective of this research is to explore how a back propagation neural network (BNN) can be applied to the speech recognition of isolated words. The simulation results show that a BNN offers an efficient approach for small vocabulary systems. The recognition rate reaches 100% for a 5 word system and 94% for a 10 word system. The general techniques developed in this article can be extended to other applications, such as the recognition of sonar targets and the classification of underwater acoustic signals.

A method for underwater acoustic signal classification using convolutional neural network combined with discrete wavelet transform

The detection and classification of underwater targets such as fish are one of the major tasks of the underwater acoustic signal processing and are very important for scientific, fisheries and ocean engineering and economic fields. The convolutional neural network (CNN) combined with the discrete wavelet transform (DWT) (namely CNN_DWT) not only reduces the data processing dimension of signals and the computational costs of the signal analysis, but also improves the performance of target detection and classification. This paper proposes a new CNN to classify the images that reflected the underwater acoustic signal in the database that is made up of the scalogram of underwater acoustic signals. Also, in order to attain greater accuracy and comparable efficiency to the spatial domain processing, we convert the data to the wavelet domain. Also, we propose a deep learning method for the classification of underwater acoustic signals using the new CNN combined with DWT. Next, through the simulation experiment, we evaluate our new method for underwater acoustic signal classification using the CNN combined with DWT, by comparing with classical methods. Comparing the proposed method to spatial domain CNN and classical methods, the experimental results reveal a substantial increment in classification accuracy and noise robustness. And the learning curves show that the proposed CNN_DWT does not generate the overfitting problem and its generalization ability is high. The proposed CNN_DWT improves the classification accuracy and convergence of underwater acoustic signals than the classical CNNs. The noise robustness of the proposed CNN_DWT is higher than those of classical CNNs and back-propagation neural networks (BPNNs) for the classification of underwater acoustic signals. Experimental results show that the classification performance of new CNN combined with DWT is higher than those of classical CNNs and BPNNs for the classification of underwater acoustic signals.

Underwater Acoustic Signal Classification Based on Sparse Time–Frequency Representation and Deep Learning

For classification of underwater acoustic signals, we propose a novel sparse anisotropic chirplet transform (ACT) to reveal fine time-frequency structures. The signal features in the form of a time-frequency map are fed into a deep convolutional neural network, referred to as a time-frequency feature network (TFFNet), which brings flexibility to signal classification. The TFFNet is based on a novel efficient feature pyramid enhancing feature (EFP) maps by aggregating the context information at different scales. To remove the gridding artifacts on enhanced feature maps, a form of aggregating transformation, a forward feature fusion, is utilized to merge the forward feature maps. Main contributions of this work are a novel sparse ACT, a TFFNet classifier, and an EFP with forward feature fusion. Experimental results demonstrate that the sparse ACT provides a high-resolution time-frequency representation of underwater signals and the TFFNet improves the classification performance compared to known networks and two machine learning methods (random forest and support vector machine with radial basis function kernel) on two real data sets, an underwater acoustic communication signal data set and whale sounds data set.

A New Low SNR Underwater Acoustic Signal Classification Method Based on Intrinsic Modal Features Maintaining Dimensionality Reduction

Abstract The classification of low signal-to-noise ratio (SNR) underwater acoustic signals in complex acoustic environments and increasingly small target radiation noise is a hot research topic.. This paper proposes a new method for signal processing—low SNR underwater acoustic signal classification method (LSUASC)—based on intrinsic modal features maintaining dimensionality reduction. Using the LSUASC method, the underwater acoustic signal was first transformed with the Hilbert-Huang Transform (HHT) and the intrinsic mode was extracted. the intrinsic mode was then transformed into a corresponding Mel-frequency cepstrum coefficient (MFCC) to form a multidimensional feature vector of the low SNR acoustic signal. Next, a semi-supervised fuzzy rough Laplacian Eigenmap (SSFRLE) method was proposed to perform manifold dimension reduction (local sparse and discrete features of underwater acoustic signals can be maintained in the dimension reduction process) and principal component analysis (PCA) was adopted in the process of dimension reduction to define the reduced dimension adaptively. Finally, Fuzzy C-Means (FCMs), which are able to classify data with weak features was adopted to cluster the signal features after dimensionality reduction. The experimental results presented here show that the LSUASC method is able to classify low SNR underwater acoustic signals with high accuracy.

Blind equalization and automatic modulation classification of underwater acoustic signals

It is useful to passively characterize the underwater acoustic communications environment for a variety of purposes, including interference avoidance and enforcement of restrictions protecting marine mammals. Automatically determining the modulation of a received waveform can permit sonar or communications operations within the same bandwidth with a minimum of collisions, and it can identify a particular system operating outside its permitted regime. The characterization system needs to both determine the modulation and an unknown, time-varying channel impulse response since the transmitter and receiver are not coordinating. In this work, we demonstrate the use of blind equalization along with Convolutional Neural Networks for automatic classification of underwater signals. Our current research focuses on classification of constant modulus signals and demonstrates an approximate 30 percent improvement in modulation classification, compared to approaches without equalization, and a significant reduction in the amount of data needed for training. We considered BPSK, QPSK, MSK, FSK and 8-PSK modulations using simplified synthetic channels simulated via MATLAB to demonstrate our results. Future work is aimed at demonstrating classification improvement using realistic channel models simulated via the Sonar Simulation Toolkit, real underwater channels gathered from data collects, and additional underwater acoustic signal types.

English

Speech interface to computer is the next big step that the technology needs to take for general users. Automatic speech recognition (ASR) will play an important role in taking technology to the people. There are numerous applications of speech recognition such as direct voice input in aircraft, data entry, speech-to-text processing, voice user interfaces such as voice dialling. ASR system can be divided into two different parts, namely feature extraction and feature recognition. In this paper we present MATLAB based feature recognition using back- propagation neural network for ASR. The objective of this research is to explore how neural networks can be employed to recognize isolated-word speech as an alternative to the traditional methodologies. The general techniques developed here can be further extended to other applications such as sonar target recognition, missile tracking and classification of underwater acoustic signals. Back-propagation neural network algorithm uses input training samples and their respective desired output values to learn to recognize specific patterns, by modifying the activation values of its nodes and weights of the links connecting its nodes. Such a trained network is later used for feature recognition in ASR systems.

Detection and classification using the Estimated Ocean Detector

We have developed the Estimated Ocean Detector, a likelihood ratio receiver with an estimator‐correlator structure, and applied it to detection and classification of underwater acoustic signals. The receiver requires that the noise probability density function (pdf) to belong to the exponential class but need not be Gaussian. A composite hypothesis is employed in order to incorporate knowledge (or predictions) of the signal parameter statistics. Previously, receiver performance was demonstrated for Gaussian noise and sinusoidal signals with known frequency and phase and whose amplitude pdfs were predicted using knowledge of the ocean environment and an acoustic propagation program. In order for the receiver to be useful operationally, it must be able to accommodate unknown signal phase and incorporate numerical estimates of noise and signal parameter pdfs. Progress toward satisfying these requirements is reported in this talk. Work supported by the Office of Naval Research Undersea Signal Processing.

Multi-wavelet detection and denoising of low-frequency chirp signals using adaptive wavelet methods

The detection and classification of underwater acoustic signals embedded in noise is a fundamental problem of interest to the signal processing community. The use of wavelet transforms is a recent development in digital signal processing which has been applied in many different areas. A particular type of wavelet is the chirplet, which includes frequency variation as well as time shift and scaling. Both linear and polynomial chirp signals are present in underwater acoustic signals generated by such sources as biologics, ships, and submarines. Distinguishing the features of these chirps relative to other ambient noise shows promise as an initial step in classification of underwater acoustic signals. Removal of unwanted broadband signal components via wavelet methods has been shown to outperform other noise removal processes such as low-pass and high-pass filtering and Weiner filtering. Examples of low-frequency simulated chirp signals with additive noise have been generated. A multi-wavelet packet method for detection and denoising low-frequency signals containing multiple chirps embedded in noise using a specific wavelet designed for polynomial chirp signals is shown. [Research supported in part by ONR.]

Wavelet detection and denoising of low-frequency chirp signals

The detection and classification of underwater acoustic signals embedded in noise is a fundamental problem of interest to the Navy. The use of wavelet transforms is a recent development in digital signal processing that has been applied in many different areas. A particular type of wavelet is the chirplet, which includes frequency variation as well as time shift and scaling. The analysis of low-frequency signals containing multiple chirps using wavelet and chirplet techniques is demonstrated. Examples of low-frequency synthetic chirp signals are generated. Denoising and feature extraction of these signals using various wavelets with wavelet packet techniques are described.

Maximum likelihood approach to image texture and acoustic signal classification

The authors describe a method of classifying natural textures based on the maximum likelihood parameter estimation technique. The novelty of the technique lies in the use of textural features that are derived from the subbands of a wavelet transformed image via the co-occurrence matrices. A maximum likelihood classifier is designed using a set of training texture samples. Ten different Brodotz textures have been classified using this procedure with an average classification accuracy of 99.7%. The main emphasis is to apply this technique to the classification of underwater acoustic signals. A time–frequency plot is obtained for each segment of the acoustic signal and then converted to an intensity pattern. The textural classification scheme is then applied to the intensity patterns of the acoustic signals. Eight different underwater acoustic signals have been classified by this procedure with an average accuracy of 99.99%.

On the classification of underwater acoustic signals. I. An environmentally adaptive approach

Environmental conditions and geometries have a pronounced effect upon signals received as scattered underwater acoustic waveforms. Information which may be processed to yield accurate assessments of the current environmental condition (and geometry) is important to the effective analysis and classification of these received signals. However, such environmental information is often ignored, or at best introduced indirectly, in degradation, through the usual formulation of composite hypothesis tests. Here, a methodology is developed which permits one to introduce estimates of the current environmental state (i.e., certain environments and object properties) in order to adapt the classification algorithm to the situation. The implications of such an environmentally adaptive approach are discussed, and general procedures for estimation, feature extraction, and feature selection are then presented. The pertinent features of analytic statistical–physical models of the medium provide useful guides, not only for the type of experiments needed to define these environmental states, but also for models, for deriving ’’artificial’’ data, or waveforms, by which the adaptive portions of the classification process can be studied and implemented.

On the classification of underwater acoustic signals. II. Experimental applications involving fish

In a previous paper (Part I) a general approach to adaptive classification has been presented. That approach is applied here (Part II) to classification of waveforms obtained in a number of illustrative experiments. Experiments at 116 kHz (in the ARL Test Tank Facility) have provided backscattered waveforms (echoes) from known types of scatterers, randomly located and with random acoustic cross sections. Scatterers of primary interest here are a variety of small fish. Signals which form competing categories are scattering from seaweed (aquarium grass), surface reverberation, and artifacts (glass bottles). The data obtained are representative of three classification problems that have some practical significance in underwater acoustics. These three problems require distinguishing between the received signals which result from scattering by: (1) fish versus air–water surface, seaweed, and/or artifacts; (2) fish and air–water surface versus air–water surface only; and (3) ’’low density’’ versus ’’medium density’’ versus ’’high density’’ of fish. In each case the waveforms are analyzed and features appropriate to the classification task are extracted from a portion of the data. These features are then used to design a classification algorithm for the environmentally adaptive approach and for its conventional counterpart. These algorithms are then used on the remaining data to make classifications. The superiority and usefulness of the proposed environmentally adaptive approach is demonstrated here by the noticeable improvement in quantitative classification which results from the exploitation of environmental state data.

On the classification of underwater acoustic signals. I. An environmentally adaptive approach

Environmental conditions and geometries have a pronounced effect upon the signals received as scattered underwater acoustic waveforms. The use of these environmental factors is important. However, these factors are often ignored or at best are introduced indirectly in the formulation of composite hypothesis testing when analyzing the received signals and classifying causes of the scattering. Here we develop a methodology which permits one to introduce estimates of the current environmental state (i.e., certain environments and geometries) in order to adapt the classification algorithm to the situation. The pertinent features of statistical-physical models of the medium (as proposed by Middleton, for example) are useful guides to the experiments needed for the definition of these environmental states. The implications of an environmentally adaptive approach are discussed, and procedures for estimation, feature extraction, and feature selection are presented. [Work supported by the Naval Sea Systems Command.]

On the classification of underwater acoustic signals. II. Experimental applications involving fish

The general technique described in Part I is applied to several experimental classification problems. An experiment conducted at 116 kHz in the ARL Test Tank Facility is described, in which it was possible to obtain echoes from a known number of scatterers random in location and acoustic cross section. The main scatterers of interest here are one variety of small fish, all of about the same length (10 cm). Other origins of backscattering, which form the signals for competing categories, are seaweed (aquarium grass), surface reverberation (mechanically driven), and an artifact (glass bottle). Using these data, we formulate three classification problems. The three problems are (generally) the categorization of received signals which result from scattering by (a) fish versus surface, seaweed, and/or artifacts; (b) fish and surface versus surface; and (c) low versus medium versus a high density of fish. Classifiers are designed and tested with and without estimates of the environmental state. The usefulness of the approach is illustrated by the improvement in performance which results from monitoring the environmental state. [Work supported by the Naval Sea Systems Command.]