Ship detection and recognition based on high-resolution remote sensing imagery

Yansong Li; Guangqing Xia; Xing Zhong; Ruifei Zhu; Chaoli Zeng

doi:10.1117/12.3009074

29 December 2023 Ship detection and recognition based on high-resolution remote sensing imagery

Yansong Li, Guangqing Xia, Xing Zhong, Ruifei Zhu, Chaoli Zeng

Proceedings Volume 12976, Eighth Asia Pacific Conference on Optics Manufacture and Third International Forum of Young Scientists on Advanced Optical Manufacturing (APCOM and YSAOM 2023); 1297617 (2023) https://doi.org/10.1117/12.3009074
Event: 8th Asia Pacific Conference on Optics Manufacture & 3rd International Forum of Young Scientists on Advanced Optical Manufacturing, 2023, Shenzhen, China

Abstract

Ship detection plays a critical role in maritime safety and surveillance. Targets in high-resolution remote sensing imagery frequently have complex textures and shapes due to the rapid advancement of remote sensing satellite technology. These features can be efficiently captured and accurately detected by deep learning models. Due to the complicated harbor environment, a variety of ship types, and a huge volume of ships, ship detection based on high-resolution remote sensing photos is still a difficult problem when compared to generic object detection. For the purpose of detecting ships in high resolution remote sensing photos, this research suggests the EBENet method. In order to increase the model's precision, the entire method employs the lightweight network EfficientNet as the YOLOv5 backbone network and integrates the BiFPN and ECA attention mechanisms. The attention mechanism module helps the model to focus more on critical features, boosting the accuracy of ship detection, while the bi-directional feature pyramid network module integrates multiscale features to enhance the precision of object recognition. Additionally, research was done utilizing publicly accessible optical remote sensing image datasets to classify different ship types into three groups: Cargo ship, Warship, and Civilian ship. Comparative tests using the YOLOv5 technique were conducted. The experimental findings show that in ship detecting tasks, EBENet beats the YOLOv5 algorithm, with a mAP50 of 93.6% and FPS of 104.189, performs better in reducing missed detections and false alarms while retaining high detection efficiency. Therefore, EBENet hence has potential for use in a variety of maritime monitoring and safety duties.

(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.

Citation Download Citation

Yansong Li, Guangqing Xia, Xing Zhong, Ruifei Zhu, and Chaoli Zeng "Ship detection and recognition based on high-resolution remote sensing imagery", Proc. SPIE 12976, Eighth Asia Pacific Conference on Optics Manufacture and Third International Forum of Young Scientists on Advanced Optical Manufacturing (APCOM and YSAOM 2023), 1297617 (29 December 2023); https://doi.org/10.1117/12.3009074

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
7 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Object detection

Remote sensing

Education and training

Ocean optics

Detection and tracking algorithms

Feature fusion

Image enhancement

Show All Keywords

Keywords/Phrases

Search In:

Publication Years