2022
DOI: 10.3390/s22062181
|View full text |Cite
|
Sign up to set email alerts
|

A Survey of Underwater Acoustic Data Classification Methods Using Deep Learning for Shoreline Surveillance

Abstract: This paper presents a comprehensive overview of current deep-learning methods for automatic object classification of underwater sonar data for shoreline surveillance, concentrating mostly on the classification of vessels from passive sonar data and the identification of objects of interest from active sonar (such as minelike objects, human figures or debris of wrecked ships). Not only is the contribution of this work to provide a systematic description of the state of the art of this field, but also to identif… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
2

Citation Types

0
9
0

Year Published

2022
2022
2024
2024

Publication Types

Select...
7
1

Relationship

0
8

Authors

Journals

citations
Cited by 34 publications
(9 citation statements)
references
References 138 publications
(186 reference statements)
0
9
0
Order By: Relevance
“…SSD comprises a backbone model, typically a pre-trained image classification network serving as a feature extractor, and an SSD head, consisting of additional convolutional layers interpreting outputs as bounding boxes and object classes. YOLO, in contrast, generates multiple bounding boxes per object and suppresses redundant boxes to obtain final coordinates [92,93].…”
Section: Deep Learningmentioning
confidence: 99%
See 1 more Smart Citation
“…SSD comprises a backbone model, typically a pre-trained image classification network serving as a feature extractor, and an SSD head, consisting of additional convolutional layers interpreting outputs as bounding boxes and object classes. YOLO, in contrast, generates multiple bounding boxes per object and suppresses redundant boxes to obtain final coordinates [92,93].…”
Section: Deep Learningmentioning
confidence: 99%
“…Subsequently, the second stage performs object classification. Examples of these architectures are the Fast R-CNN, Faster R-CNN, and Mask R-CNN [93,94].…”
Section: Deep Learningmentioning
confidence: 99%
“…In recent years, many methods have attempted to establish end-to-end deep neural networks to identify underwater acoustic targets by automatically extracting deep features 1 . However, because of expensive annotation cost and inevitable annotation errors 2 , there is a lack of massive high-quality labeled samples to train robust deep neural networks in underwater acoustic target recognition 3 .…”
Section: Introductionmentioning
confidence: 99%
“…The (2, 1)-norm favors a small number of nonzero rows in the matrix W , therefore ensuring that the common features (most effective centers) will be selected. It should be noted that, Regularization techniques [10,11] proved to improve the generalization ability and therefore the performance of a model. A comprehensive study and a state-of-the-art review of the regularization strategies in machine learning is given in [10].…”
Section: Introductionmentioning
confidence: 99%
“…A comprehensive study and a state-of-the-art review of the regularization strategies in machine learning is given in [10]. It is being used in different classification problems such as, image recognition [12], Underwater Acoustic Data Classification [11] e.t.c. The above two approaches MDS and RMDS were for unsupervised data.…”
Section: Introductionmentioning
confidence: 99%