Ship video detection based on improved YOLOv4-Tiny

Liuling Kong; Xiuwen Liu

doi:10.1117/12.2642851

28 July 2022 Ship video detection based on improved YOLOv4-Tiny

Liuling Kong, Xiuwen Liu

Proceedings Volume 12303, International Conference on Cloud Computing, Internet of Things, and Computer Applications (CICA 2022); 1230326 (2022) https://doi.org/10.1117/12.2642851
Event: International Conference on Cloud Computing, Internet of Things, and Computer Applications, 2022, Luoyang, China

Abstract

Aiming at the problems of target loss and poor real-time performance in ship detection at sea, a target detection method based on improved YOLOv4-Tiny network was proposed to realize real-time detection of ship video. Firstly, an output prediction scale of 52×52 is added to the network structure to improve the detection ability of small targets on ships. Secondly, based on the improved model structure and self-established data set, the priori anchor frame was redesigned. Finally, the mish activation function is used to improve the model performance of YOLOv4-Tiny. The experimental results show that the improved YOLOv4-Tiny algorithm not only retains the great advantage of fast detection speed, but also significantly improves the detection accuracy. It can realize real-time high-precision detection of ship video, which is of great significance for maritime safety law enforcement and supervision.

Citation Download Citation

Liuling Kong and Xiuwen Liu "Ship video detection based on improved YOLOv4-Tiny", Proc. SPIE 12303, International Conference on Cloud Computing, Internet of Things, and Computer Applications (CICA 2022), 1230326 (28 July 2022); https://doi.org/10.1117/12.2642851

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