Remote sensing object detection based on  lightweight YOLO-v4

Keng Li; Yunfei Cao; He Chen

doi:10.1117/12.2639675

19 October 2022 Remote sensing object detection based on lightweight YOLO-v4

Keng Li, Yunfei Cao, He Chen

Proceedings Volume 12294, 7th International Symposium on Advances in Electrical, Electronics, and Computer Engineering; 122944O (2022) https://doi.org/10.1117/12.2639675
Event: 7th International Symposium on Advances in Electrical, Electronics and Computer Engineering (ISAEECE 2022), 2022, Xishuangbanna, China

Abstract

The remote sensing object detection algorithms based on deep learning have high detection performances, but the network structures are too complexity to meet the real-time processing requirements in on-board remote sensing object detection. In order to solve this problem, we proposed a lightweight YOLO-v4 network, which is 76% smaller than the original YOLO-v4. As for the decrease of lightweight network’s accuracy, we adopted the general instance distillation algorithm, which used the original YOLO-v4 network as the teacher network and whose detection accuracy achieved 2.1% mAP gain.

Citation Download Citation

Keng Li, Yunfei Cao, and He Chen "Remote sensing object detection based on lightweight YOLO-v4", Proc. SPIE 12294, 7th International Symposium on Advances in Electrical, Electronics, and Computer Engineering, 122944O (19 October 2022); https://doi.org/10.1117/12.2639675

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