Optical resolution requirements for effective deep learning-based ship recognition

Zhenduo Zhang; Shihao Liu; Ming Xie; Ying Li; Huijie Wang

doi:10.1117/12.2690113

7 September 2023 Optical resolution requirements for effective deep learning-based ship recognition

Zhenduo Zhang, Shihao Liu, Ming Xie, Ying Li, Huijie Wang

Proceedings Volume 12790, Eighth International Conference on Electromechanical Control Technology and Transportation (ICECTT 2023); 127904I (2023) https://doi.org/10.1117/12.2690113
Event: 8th International Conference on Electromechanical Control Technology and Transportation (ICECTT 2023), 2023, Hangzhou, China

Abstract

The number of ships in the maritime cargo transportation industry has increased significantly. To ensure ship safety, deep learning can be used to realize ship recognition and positioning. However, the current research focuses on improving recognition accuracy, and the hardware requirements of related systems have not been investigated extensively. Thus, in this paper, we propose a deep learning-based recognition model that demonstrates high average accuracy, and we explore the minimum pixel requirements that this network can recognize. The proposed model employs the YOLOv4 network combined with K-means clustering method as the backbone network. Targeting the minimum pixel problem of ship targets that can be identified by the proposed model, we consider multiple groups of downsampling convolution check test data of different sizes to downsample to different degrees, record the average accuracy of different downsampled data input to the network, and discuss the relationship between the size and distance of targets and imaging resolution. The test results were tested under various conditions, and the minimum pixel requirements for ship recognition were calculated for an ideal optical system, and the optimal settings of an onboard optical system were determined.

Citation Download Citation

Zhenduo Zhang, Shihao Liu, Ming Xie, Ying Li, and Huijie Wang "Optical resolution requirements for effective deep learning-based ship recognition", Proc. SPIE 12790, Eighth International Conference on Electromechanical Control Technology and Transportation (ICECTT 2023), 127904I (7 September 2023); https://doi.org/10.1117/12.2690113

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
11 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Target detection

Object detection

Convolution

Detection and tracking algorithms

Education and training

RGB color model

Target recognition

Show All Keywords

Keywords/Phrases

Search In:

Publication Years