Remote Sensing Aircraft Target Detection Algorithm Based on Improved YOLOv5s
(2024) 5th International Seminar on Artificial Intelligence, Networking and Information Technology, AINIT 2024 In 2024 5th International Seminar on Artificial Intelligence, Networking and Information Technology, AINIT 2024 p.681-686- Abstract
To solve the problem of low precision measurement accuracy of aircraft target algorithm in remote sensing images, an improved YOLOv5s algorithm was proposed based on YOLOv5s algorithm for aircraft target identification. The backbone network, neck network, output and loss function parts of YOLOv5s network were improved respectively. Spatial pyramid pooling-fast module was replaced by atrous spatial pyramid pooling module in the backbone network, conventional convolution was replaced by deformable convolution in the neck network, convolutional block attention module was added in the output, and complete intersection over union loss function was replaced by SCYLLA intersection over union loss function. On the MAR20 dataset, the experiments... (More)
To solve the problem of low precision measurement accuracy of aircraft target algorithm in remote sensing images, an improved YOLOv5s algorithm was proposed based on YOLOv5s algorithm for aircraft target identification. The backbone network, neck network, output and loss function parts of YOLOv5s network were improved respectively. Spatial pyramid pooling-fast module was replaced by atrous spatial pyramid pooling module in the backbone network, conventional convolution was replaced by deformable convolution in the neck network, convolutional block attention module was added in the output, and complete intersection over union loss function was replaced by SCYLLA intersection over union loss function. On the MAR20 dataset, the experiments showed that the accuracy of the improved YOLOv5s algorithm was 89.5%, which was 1.6% higher than that of the YOLOv5s algorithm.
(Less)
- author
- Yang, Xiao and Xiao, Nanyu
- publishing date
- 2024
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- keywords
- atrous spatial pyramid pooling, convolutional block attention module, deformable convolution, remote sensing, SCYLLA intersection over union, YOLOv5s
- host publication
- 2024 5th International Seminar on Artificial Intelligence, Networking and Information Technology, AINIT 2024
- series title
- 2024 5th International Seminar on Artificial Intelligence, Networking and Information Technology, AINIT 2024
- pages
- 6 pages
- publisher
- IEEE - Institute of Electrical and Electronics Engineers Inc.
- conference name
- 5th International Seminar on Artificial Intelligence, Networking and Information Technology, AINIT 2024
- conference location
- Hybrid, Nanjing, China
- conference dates
- 2024-05-29 - 2024-05-31
- external identifiers
-
- scopus:85199172923
- ISBN
- 9798350385557
- DOI
- 10.1109/AINIT61980.2024.10581820
- language
- English
- LU publication?
- no
- additional info
- Publisher Copyright: © 2024 IEEE.
- id
- c74e1481-99b6-4d7d-8206-ae690fec5435
- date added to LUP
- 2024-12-04 10:31:28
- date last changed
- 2025-04-04 14:22:36
@inproceedings{c74e1481-99b6-4d7d-8206-ae690fec5435,
abstract = {{<p>To solve the problem of low precision measurement accuracy of aircraft target algorithm in remote sensing images, an improved YOLOv5s algorithm was proposed based on YOLOv5s algorithm for aircraft target identification. The backbone network, neck network, output and loss function parts of YOLOv5s network were improved respectively. Spatial pyramid pooling-fast module was replaced by atrous spatial pyramid pooling module in the backbone network, conventional convolution was replaced by deformable convolution in the neck network, convolutional block attention module was added in the output, and complete intersection over union loss function was replaced by SCYLLA intersection over union loss function. On the MAR20 dataset, the experiments showed that the accuracy of the improved YOLOv5s algorithm was 89.5%, which was 1.6% higher than that of the YOLOv5s algorithm.</p>}},
author = {{Yang, Xiao and Xiao, Nanyu}},
booktitle = {{2024 5th International Seminar on Artificial Intelligence, Networking and Information Technology, AINIT 2024}},
isbn = {{9798350385557}},
keywords = {{atrous spatial pyramid pooling; convolutional block attention module; deformable convolution; remote sensing; SCYLLA intersection over union; YOLOv5s}},
language = {{eng}},
pages = {{681--686}},
publisher = {{IEEE - Institute of Electrical and Electronics Engineers Inc.}},
series = {{2024 5th International Seminar on Artificial Intelligence, Networking and Information Technology, AINIT 2024}},
title = {{Remote Sensing Aircraft Target Detection Algorithm Based on Improved YOLOv5s}},
url = {{http://dx.doi.org/10.1109/AINIT61980.2024.10581820}},
doi = {{10.1109/AINIT61980.2024.10581820}},
year = {{2024}},
}