Abstract
For the purpose of improving the accuracy of underwater acoustic target recognition with only a small number of labeled data, we proposed a novel recognition method, including 4 steps: pre-processing, pre-training, fine-tuning and recognition. The 4 steps can be explained as follows: (1) Pre-processing with Resonance-based Sparsity Signal Decomposition (RSSD): RSSD was firstly utilized to extract high-resonance components from ship-radiated noise. The high-resonance components contain the major information for target recognition. (2) Pre-training with unsupervised feature-extraction: we proposed a one-dimensional convolution autoencoder-decoder model and then we pre-trained the model to extract features from the high-resonance components. (3) Fine-tuning with supervised feature-separation: a supervised feature-separation algorithm was proposed to fine-tune the model and separate the extracted features. (4) Recognition: classifiers were trained to recognize the separated features and complete the recognition mission. The unsupervised pre-training autoencoder-decoder can make good use of a large number of unlabeled data, so that only a small number of labeled data are required in the following supervised fine-tuning and recognition, which is quite effective when it is difficult to collect enough labeled data. The recognition experiments were all conducted on ship-radiated noise data recorded using a sensory hydrophone. By combining the 4 steps above, the proposed recognition method can achieve recognition accuracy of 93.28%, which sufficiently surpasses other traditional state-of-art feature-extraction methods.
Highlights
When a ship moves in the water, it produces noise, called ship-radiated noise
Underwater acoustic target recognition depends on the decisions of well-trained sonar men, which can be highly inaccurate due to the need of continuous monitoring, and at times much affected by weather conditions
The original and main contributions of this paper can be briefly summarized as follows: (1) we proposed a novel recognition method of 4 steps for underwater acoustic target recognition; (2) we carefully considered the oscillatory nature of ship-radiated noise and we proved the effectiveness of Resonance-based Sparsity Signal Decomposition (RSSD) used in extracting “invariant” part of the ship-radiated noise for recognition; (3) we specially designed a totally new model with specific structures for extracting informative and invariant features for underwater acoustic target recognition; (4) we created a totally new universal loss function which named “feature-separation” algorithm for recognition
Summary
When a ship moves in the water, it produces noise, called ship-radiated noise. Due to the unique characteristics of the radiated noise of different classes of ships, it is possible to identify a specific class of ships or even a specific ship by analyzing the ship-radiated noise. Recognition for ship-radiated underwater noise is one of the most important and challenging subjects in underwater acoustic signal processing. Underwater acoustic target recognition depends on the decisions of well-trained sonar men, which can be highly inaccurate due to the need of continuous monitoring, and at times much affected by weather conditions. Underwater acoustic target recognition is a complex pattern recognition problem. Due to the difficulty in collecting a large number of ship-radiated noise data, target recognition from ship-radiated noise is typically done under limited samples or even small samples
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have