Lightweight Convolutional Neural Network for False Alarm Elimination in SAR Ship Detection

Xingfei Zheng; Yangkai Feng; Hao Shi; Bocheng Zhang; Liang Chen

doi:10.1049/icp.2021.0801

Lightweight Convolutional Neural Network for False Alarm Elimination in SAR Ship Detection

Xingfei Zheng
, Yangkai Feng
, Hao Shi^*
, Bocheng Zhang
, Liang Chen

^*Corresponding author for this work

School of Information and Electronics

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

5 Citations (Scopus)

Abstract

SAR ship detection is an important task of SAR image interpretation and plays a significant role in global marine surveillance. However, since there are usually many false alarms in the detection results, a desirable performance is rarely achieved. Aiming at this problem, this paper proposes an improved method for the ship detection process, using the lightweight convolutional neural network EfficientNet-B0 as a false alarm elimination network, cascaded after the detection stage, to reduce the false alarm rate in the detection results and improve the overall detection accuracy. The results on the MSTAR data set show that EfficientNet-B0 has a good classification effect, and the experiments on the custom data set also verify the effectiveness of the network.

Original language	English
Title of host publication	IET Conference Proceedings
Publisher	Institution of Engineering and Technology
Pages	287-291
Number of pages	5
Volume	2020
Edition	9
ISBN (Electronic)	9781839535406
DOIs	https://doi.org/10.1049/icp.2021.0801
Publication status	Published - 2020
Event	5th IET International Radar Conference, IET IRC 2020 - Virtual, Online Duration: 4 Nov 2020 → 6 Nov 2020

Conference

Conference	5th IET International Radar Conference, IET IRC 2020
City	Virtual, Online
Period	4/11/20 → 6/11/20

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 14 Life Below Water

Keywords

False alarm elimination
Ship detection
Synthetic aperture radar

Access to Document

10.1049/icp.2021.0801

Cite this

@inproceedings{5a0f02e5158d4c7481a0f5b57e55a412,

title = "Lightweight Convolutional Neural Network for False Alarm Elimination in SAR Ship Detection",

abstract = "SAR ship detection is an important task of SAR image interpretation and plays a significant role in global marine surveillance. However, since there are usually many false alarms in the detection results, a desirable performance is rarely achieved. Aiming at this problem, this paper proposes an improved method for the ship detection process, using the lightweight convolutional neural network EfficientNet-B0 as a false alarm elimination network, cascaded after the detection stage, to reduce the false alarm rate in the detection results and improve the overall detection accuracy. The results on the MSTAR data set show that EfficientNet-B0 has a good classification effect, and the experiments on the custom data set also verify the effectiveness of the network.",

keywords = "False alarm elimination, Ship detection, Synthetic aperture radar",

author = "Xingfei Zheng and Yangkai Feng and Hao Shi and Bocheng Zhang and Liang Chen",

note = "Publisher Copyright: {\textcopyright} 2020 IET Conference Proceedings. All rights reserved.; 5th IET International Radar Conference, IET IRC 2020 ; Conference date: 04-11-2020 Through 06-11-2020",

year = "2020",

doi = "10.1049/icp.2021.0801",

language = "English",

volume = "2020",

pages = "287--291",

booktitle = "IET Conference Proceedings",

publisher = "Institution of Engineering and Technology",

address = "United Kingdom",

edition = "9",

}

TY - GEN

T1 - Lightweight Convolutional Neural Network for False Alarm Elimination in SAR Ship Detection

AU - Zheng, Xingfei

AU - Feng, Yangkai

AU - Shi, Hao

AU - Zhang, Bocheng

AU - Chen, Liang

PY - 2020

Y1 - 2020

N2 - SAR ship detection is an important task of SAR image interpretation and plays a significant role in global marine surveillance. However, since there are usually many false alarms in the detection results, a desirable performance is rarely achieved. Aiming at this problem, this paper proposes an improved method for the ship detection process, using the lightweight convolutional neural network EfficientNet-B0 as a false alarm elimination network, cascaded after the detection stage, to reduce the false alarm rate in the detection results and improve the overall detection accuracy. The results on the MSTAR data set show that EfficientNet-B0 has a good classification effect, and the experiments on the custom data set also verify the effectiveness of the network.

AB - SAR ship detection is an important task of SAR image interpretation and plays a significant role in global marine surveillance. However, since there are usually many false alarms in the detection results, a desirable performance is rarely achieved. Aiming at this problem, this paper proposes an improved method for the ship detection process, using the lightweight convolutional neural network EfficientNet-B0 as a false alarm elimination network, cascaded after the detection stage, to reduce the false alarm rate in the detection results and improve the overall detection accuracy. The results on the MSTAR data set show that EfficientNet-B0 has a good classification effect, and the experiments on the custom data set also verify the effectiveness of the network.

KW - False alarm elimination

KW - Ship detection

KW - Synthetic aperture radar

UR - https://www.scopus.com/pages/publications/85140016222

U2 - 10.1049/icp.2021.0801

DO - 10.1049/icp.2021.0801

M3 - Conference contribution

AN - SCOPUS:85140016222

VL - 2020

SP - 287

EP - 291

BT - IET Conference Proceedings

PB - Institution of Engineering and Technology

T2 - 5th IET International Radar Conference, IET IRC 2020

Y2 - 4 November 2020 through 6 November 2020

ER -

Lightweight Convolutional Neural Network for False Alarm Elimination in SAR Ship Detection

Abstract

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this