Journal of Applied Optics, Volume. 45, Issue 5, 975(2024)
Infrared dim target detection method based on improved YOLOv5s model
Figures & Tables(11)
Fig. 8. Comparison of target detection results before and after YOLOv5s improvement
Table 1. Configuration of experimental platform
View tableView in ArticleTable 1. Configuration of experimental platform
名称 环境参数 操作系统 Ubuntu CPU Intel(R) Xeon(R) Platinum 8255C GPU NIVDIA RTX 2080 Ti深度学习框架 Pytorch1.5.1 CUDA版本 10.1
Table 2. Ablation experimental results
View tableView in ArticleTable 2. Ablation experimental results
组号 Precision/% Recall/% mAP@0.5/% FPS/(frame·s−1) FLOPS/G Params/M YOLOv5s 93.55 92.27 90.04 73 17.303 6.627 YOLOv5s中引入TCC 94.91 92.96 91.91 70 17.044 6.303 YOLOv5s中引入Ghost 95.11 93.64 91.94 64 16.520 6.288 YOLOv5s中引入TCC+Ghost 94.89 92.61 92.30 68 16.260 5.964
Tools
Get Citation
Copy Citation Text
Jianjun ZHANG, Yudan CHEN, Yuling LIU, Mingming ZHANG, Fuyu HUANG. Infrared dim target detection method based on improved YOLOv5s model[J]. Journal of Applied Optics, 2024, 45(5): 975
Paper Information
Category:
Received: Mar. 14, 2024
Accepted: --
Published Online: Dec. 20, 2024
The Author Email: HUANG Fuyu (黄富瑜)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20