Spacecraft Recovery & Remote Sensing, Volume. 46, Issue 1, 123(2025)
Object Detection in Remote Sensing Images Based on Super-Resolution Hierarchical Fusion Mechanism
Figures & Tables(14)
![[in Chinese]](/Images/icon/loading.gif){kind=icon}
Table 1. VEDAI benchmark detection performance
View tableView in ArticleTable 1. VEDAI benchmark detection performance
方法 模态 各分类目标平均精度AP/% mAP50/% 参数量 FLOPs 轿车 皮卡 露营车 货车 其他 拖拉机 汽艇 面包车 YOLOv8s RGB 78.91 65.29 65.98 54.14 44.46 58.71 31.48 48.97 55.99 11.13×106 9.1×109 IR 77.05 66.98 68.43 57.90 27.29 40.02 29.76 57.26 53.09 11.13×106 9.1×109 Multi 77.38 62.39 67.15 53.01 45.28 58.86 32.94 55.54 56.57 11.13×106 9.2×109
Table 2. Ablation experiment
View tableView in ArticleTable 2. Ablation experiment
实验 残差融合 单分支增强 超分辅助分支 L1 损失 mAP50/% 参数量 FLOPs 注：√ 表示启用模块或改动。 A √ 70.56 12.49×106 29.6×109 B √ √ 72.94 12.75×106 30.3×109 C √ √ √ 78.42 13.25×106 31.0×109 D √ √ √ √ 79.45 13.25×106 31.2×109
Table 3. VEDAI accuracy comparison table
View tableView in ArticleTable 3. VEDAI accuracy comparison table
方法 模态 各分类目标平均精度AP/% mAP50/% 参数量 FLOPs 轿车 皮卡 露营车 货车 其他 拖拉机 汽艇 面包车 YOLOv8s[ 24 ]RGB 78.91 65.29 65.98 54.14 44.46 58.71 31.48 48.97 55.99 11.13×106 9.1×109 IR 77.05 66.98 68.43 57.90 27.29 40.02 29.76 57.26 53.09 11.13×106 9.1×109 Multi 77.38 62.39 67.15 53.01 45.28 58.86 32.94 55.54 56.57 11.13×106 9.2×109 Dual-YOLO[ 23 ]Multi 80.07 68.01 66.12 51.52 45.76 64.38 21.62 40.93 54.82 12.19×106 19.3×109 DM-YOLO[ 25 ]RGB 79.68 74.49 77.09 80.97 37.33 70.65 60.84 63.56 68.83 12.37×106 20.4×109 IR 76.77 74.35 64.74 63.45 45.04 78.12 70.04 77.91 68.18 12.37×106 20.4×109 Multi 86.71 75.98 66.62 77.17 43.04 62.30 70.75 84.38 70.87 12.38×106 20.5×109 SuperYOLO[ 19 ]RGB 90.30 82.66 76.69 68.55 53.86 79.48 58.08 70.30 72.49 10.83×106 26.5×109 IR 87.90 81.39 76.90 61.56 39.39 60.56 46.08 71.00 65.60 10.83×106 26.5×109 Multi 91.13 85.66 79.30 70.18 57.33 80.41 60.24 76.50 75.09 10.85×106 29.6×109 本文方法 Multi 92.04 88.98 84.98 75.25 63.39 84.67 69.24 77.05 79.45 13.25×106 31.2×109
Table 4. Drone Vehicle accuracy comparison table
View tableView in ArticleTable 4. Drone Vehicle accuracy comparison table
方法 模态 各分类目标平均精度AP/% mAP50/% 参数量 FLOPs 轿车 货车 卡车 公交车 面包车 YOLOv8s[ 24 ]RGB 90.89 43.91 58.87 89.76 43.84 65.45 11.13×106 14.2×109 IR 91.03 48.82 59.71 92.79 50.92 68.65 11.13×106 14.2×109 Multi 94.59 54.88 66.14 90.51 58.17 72.86 11.13×106 14.3×109 Dual-YOLO[ 23 ]Multi 98.16 52.87 65.72 95.77 46.58 71.82 12.19×106 10.5×109 DM-YOLO[ 25 ]RGB 89.95 48.26 58.27 89.65 44.43 66.11 12.37×106 22.6×109 IR 97.63 62.33 72.54 93.74 53.94 76.04 12.37×106 22.6×109 Multi 97.91 64.34 74.63 95.05 57.74 77.93 12.38×106 22.7×109 SuperYOLO[ 19 ]RGB 86.82 32.53 49.13 82.88 36.68 57.61 10.83×106 25.9×109 IR 97.35 53.25 65.61 92.30 49.13 71.53 10.83×106 25.9×109 Multi 97.49 56.05 69.11 93.61 51.79 73.61 10.85×106 28.1×109 本文方法 Multi 98.17 68.61 81.68 96.00 61.98 81.29 13.25×106 26.1×109
Tools
Get Citation
Copy Citation Text
Lingfeng YIN, Xudong TONG, Huan NI. Object Detection in Remote Sensing Images Based on Super-Resolution Hierarchical Fusion Mechanism[J]. Spacecraft Recovery & Remote Sensing, 2025, 46(1): 123
Paper Information
Category:
Received: Oct. 17, 2024
Accepted: --
Published Online: Apr. 2, 2025
The Author Email: Xudong TONG (003445@nuist.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20