Performance comparison of CNN-based deep learning models for maritime object identification using hyperspectral image

Dongmin Seo; Sangwoo Oh; Sekil Park

doi:10.1117/12.3017378

6 June 2024 Performance comparison of CNN-based deep learning models for maritime object identification using hyperspectral image

Dongmin Seo, Sangwoo Oh, Sekil Park

Author Affiliations +

Proceedings Volume 13024, Optical Instrument Science, Technology, and Applications III; 130240I (2024) https://doi.org/10.1117/12.3017378
Event: SPIE Optical Systems Design, 2024, Strasbourg, France

Conference Poster

Abstract

Spectrum data obtained from hyperspectral optical systems were analyzed with a CNN-based deep learning model to detect and identify maritime small objects. The hyperspectral data set for learning was extracted from more than 60 aerial observation images, and classification accuracy was derived by applying a total of 7 CNN models. Among the models used, Inception_v3 was the best at 94.9%, and this result showed more than 10% improvement in accuracy over previous studies conducted with multi-layer perceptron (MLP). If further research breaks down classification items and increases the size of datasets, we expect that the technology will become increasingly utilized in the field of maritime search and surveillance.

Citation Download Citation

Dongmin Seo, Sangwoo Oh, and Sekil Park "Performance comparison of CNN-based deep learning models for maritime object identification using hyperspectral image", Proc. SPIE 13024, Optical Instrument Science, Technology, and Applications III, 130240I (6 June 2024); https://doi.org/10.1117/12.3017378

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