DCP-YOLO：轻量化反无人机目标检测网络

刘佳; 毛碟; 蒋子瑞; 曹一涵; 陈大鹏

doi:10.11999/JEIT261031

DCP-YOLO：轻量化反无人机目标检测网络

doi: 10.11999/JEIT261031 cstr: 32379.14.JEIT261031

刘佳^{1, 2, 3, 4,},
毛碟^{1, 2},
蒋子瑞^{1, 2},
曹一涵^{1, 2, 3, 4},
陈大鹏^{1, 2, 3, 4, ,}

1.
南京信息工程大学自动化学院南京 210044
2.
江苏省智能气象探测机器人工程研究中心南京 210044
3.
江苏省大气环境与装备技术协同创新中心南京 210044
4.
大数据分析与智能系统江苏省高校重点实验室南京 210044

基金项目: 国家自然科学基金(项目编号：62473200和62203226)，江苏省青年科技人才支持计划(项目编号：JSTJ-2024-195)，江苏省“青蓝工程”

详细信息

作者简介:
刘佳：女，教授，研究方向为计算机视觉与图像处理、智能机器人，邮箱　liujia@nuist.edu.cn

毛碟：女，硕士生，研究方向为计算机视觉与图像处理、智能机器人

蒋子瑞：男，硕士生，研究方向为多模态智能体与具身智能

曹一涵：男，副教授，研究方向为红外探测与融合、自动化检测系统

陈大鹏：男，教授，研究方向为多模态信息处理、具身智能，邮箱　dpchen@nuist.edu.cn

通讯作者:
陈大鹏　dpchen@nuist.edu.cn

中图分类号: TP391.41; TN911.73; TP183
计量
- 文章访问数: 47
- HTML全文浏览量: 4
- PDF下载量: 4
- 被引次数: 0
出版历程
- 修回日期: 2026-10-19
- 录用日期: 2026-10-19
- 网络出版日期: 2026-10-28

DCP-YOLO: A Lightweight Anti-UAV Object Detection Network

LIU Jia^{1, 2, 3, 4
,},
MAO Die^{1, 2},
JIANG Zirui^{1, 2},
CAO Yihan^{1, 2, 3, 4},
CHEN Dapeng^{1, 2, 3, 4
, ,}

1.
School of Automation, Nanjing University of Information Science & Technology, Nanjing 210044, China
2.
Jiangsu Province Engineering Research Center of Intelligent Meteorological Exploration Robot, Nanjing 210044, China
3.
Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, Nanjing 210044, China
4.
Jiangsu Key Laboratory of Big Data Analysis and Intelligent Systems, Nanjing 210044, China

Funds: National Natural Science Foundation of China (62473200 and 62203226), Jiangsu Province Youth Science and Technology Talent Support Project (JSTJ-2024-195), “Qinglan Project” of Jiangsu Province

摘要

摘要: 随着低空经济的快速发展，无人机违规飞行带来严重安全隐患。针对现有反无人机检测方法存在小目标特征易丢失、边缘算力受限，难以平衡检测精度与模型轻量化的问题，该文提出一种基于YOLOv26n改进的轻量化多模块协同优化检测方法(DCP-YOLO)。首先设计DPC3k2轻量化特征提取结构，增强小目标多方向特征提取能力并缩减参数。然后，提出反向上采样模块CUS，恢复目标高频纹理细节。最后，构建C2PSCAA注意力增强模块，提升模型多尺度特征学习能力。在多个公开数据集的实验结果表明，该方法参数量仅2.29M，推理速度达210.71 FPS，检测精度较基线模型显著提升。消融与泛化实验验证了各模块的有效性与模型的泛化能力，可有效兼顾检测精度与轻量化性能，适配资源受限场景下的低空无人机实时检测任务。
- 小目标检测 /
- 反无人机检测 /
- YOLOv26n /
- 轻量化目标检测
Abstract: Objective With the rapid development of the low-altitude economy, unmanned aerial vehicles (UAVs) have been widely used in aerial photography, logistics, environmental monitoring, and emergency rescue. However, unauthorized UAV operations pose increasing challenges to public safety and low-altitude airspace management. Vision-based anti-UAV detection has attracted considerable attention due to its relatively low cost, rich visual information, and convenient deployment. Compared with conventional objects, UAV targets in low-altitude surveillance images often exhibit small scales, weak textures, significant pose variations, and complex background interference. During repeated feature downsampling, the edge and structural information of these targets may be weakened or lost, resulting in missed detections, false alarms, and inaccurate localization. Furthermore, practical anti-UAV detection systems must balance detection accuracy, computational complexity, and real-time performance under limited hardware resources. Existing lightweight detection networks may struggle to preserve discriminative spatial details while maintaining computational efficiency, particularly when UAV targets occupy only a small proportion of the input image. To address these challenges, a lightweight anti-UAV object detection network named DCP-YOLO is proposed based on YOLOv26n. The objective is to enhance the feature representation and detection accuracy of tiny UAV targets while maintaining a compact model structure and real-time inference capability. Methods The proposed DCP-YOLO is developed based on the lightweight YOLOv26n framework to improve the detection performance of tiny UAV targets while maintaining real-time inference. The network consists of three collaboratively optimized components. In the backbone network, a lightweight feature extraction module named DPC3k2 is designed by integrating depthwise separable convolution (DSConv) and pinwheel-shaped convolution (PConv) to enhance multi-directional feature representation with reduced computational complexity, enabling more discriminative feature extraction for small UAV targets. DSConv reduces redundant spatial computation through the separation of spatial feature extraction and channel mapping, while PConv employs asymmetric convolutional branches to capture directional edge and contour information. During feature fusion, a learnable reverse upsampling module named CUS replaces the conventional interpolation-based upsampling operation to recover high-frequency edge and texture information, thereby improving feature reconstruction quality. In the neck network, a contextual attention enhancement module named C2PSCAA is constructed by incorporating Context Anchor Attention (CAA) into the C2PSA structure to strengthen long-range spatial dependency modeling and multi-scale contextual representation with only a slight computational overhead. Through the collaborative optimization of lightweight feature extraction, feature reconstruction, and contextual attention modeling, the proposed network effectively enhances the representation capability of tiny UAV targets while maintaining low computational complexity and high inference efficiency. Results and Discussions Extensive experiments are conducted on the DUT Anti-UAV, TIB-Net, and VisDrone2019 datasets. The ablation results in Tables 2 and 4 demonstrate that the proposed DPC3k2, CUS, and C2PSCAA modules contribute to lightweight feature extraction, high-frequency detail reconstruction, and contextual feature modeling, respectively. On the DUT Anti-UAV dataset, their collaborative integration improves Precision, Recall, mAP@50, and mAP@50-95 to 98.7%, 90.0%, 92.4%, and 61.7%, respectively, while reducing the number of parameters to 2.29 M and maintaining an inference speed of 210.71 FPS. Similar performance improvements are observed on the TIB-Net dataset, confirming the effectiveness and complementarity of the three modules under different target scales and background conditions. The visualization results in Fig. 6 further show that the proposed method reduces missed detections and false alarms for tiny UAV targets in cluttered scenes. The comparison results in Table 5 indicate that DCP-YOLO outperforms representative YOLO-series detectors in detection accuracy while maintaining lower model complexity and a smaller model size. Compared with the baseline YOLOv26n, mAP@50 and mAP@50-95 are improved from 90.0% and 59.4% to 92.4% and 61.7%, respectively, whereas the computational complexity is reduced from 5.2 GFLOPs to 5.0 GFLOPs. The convergence curves in Fig. 8 show that DCP-YOLO maintains higher Precision and mAP@50-95 during training. Furthermore, Table 6 demonstrates that the proposed method achieves competitive or superior performance compared with recent anti-UAV and small-object detection methods while preserving a lightweight architecture. The detection results in Fig. 7 visually confirm its stronger ability to identify tiny UAV targets under complex backgrounds. The cross-dataset results in Tables 7 and 8 show that DCP-YOLO also achieves improved detection performance on VisDrone2019 while reducing the number of parameters, computational complexity, and model size. Although VisDrone2019 differs substantially from anti-UAV datasets in target categories and imaging viewpoints, the proposed method still improves mAP@50-95 by 1.1% over YOLOv26n, demonstrating favorable cross-scene adaptability and generalization capability. To investigate practical deployment feasibility, the models are further evaluated on an NVIDIA Jetson Orin Nano platform using TensorRT acceleration under a 25 W power setting, as illustrated in Fig. 9. As shown in Table 9, DCP-YOLO achieves an mAP@50 of 94.59%, an mAP@50-95 of 62.70%, and an inference speed of 116.02 FPS in FP16 mode, with a memory consumption of 3.00 GB. Although its inference speed is lower than that of the baseline on the edge platform, DCP-YOLO maintains real-time inference capability while achieving higher detection accuracy. These findings indicate that reductions in theoretical computational complexity do not necessarily translate into proportional improvements in actual inference speed, particularly when frequency-domain operations are involved. Conclusions A lightweight anti-UAV object detection network named DCP-YOLO is proposed to address the challenges of weak feature representation, high-frequency detail degradation, and limited computational resources in low-altitude UAV detection. By jointly optimizing lightweight feature extraction, learnable upsampling, and contextual attention modeling, the proposed method enhances the detection accuracy of tiny UAV targets while maintaining a compact network structure. Experimental results on DUT Anti-UAV, TIB-Net, and VisDrone2019 demonstrate the effectiveness of the proposed modules and their applicability under different scene conditions. Compared with the baseline network, DCP-YOLO achieves higher detection accuracy with fewer parameters and lower theoretical computational complexity. Edge deployment experiments further demonstrate that the proposed method can maintain real-time inference capability on a resource-constrained platform, despite the additional computational overhead associated with its feature reconstruction operations. Nevertheless, the detection of extremely small targets under severe occlusion, adverse weather, and low-light conditions remains challenging. In addition, the current network primarily relies on single-frame visual information and does not explicitly exploit temporal relationships between consecutive frames. Future work will investigate the integration of target tracking, temporal feature modeling, and multimodal perception, together with further optimization of computationally intensive operations, to improve detection robustness and deployment efficiency in complex low-altitude environments.
- Small object detection /
- Anti-UAV detection /
- YOLOv26n /
- Lightweight object detection

HTML全文

图 1 DCP-YOLO网络架构

下载: 全尺寸图片幻灯片

图 2 DPBottleneck与DPC3k2模块结构

下载: 全尺寸图片幻灯片

图 3 DSConv结构图

下载: 全尺寸图片幻灯片

图 4 PConv结构图

下载: 全尺寸图片幻灯片

图 5 C2PSCAA结构

下载: 全尺寸图片幻灯片

图 6 基于DUT Anti-UAV数据集消融实验结果可视化

下载: 全尺寸图片幻灯片

图 7 基于DUT Anti-UAV数据集YOLO系列检测结果可视化

下载: 全尺寸图片幻灯片

图 8 不同YOLO系列模型Precision和mAP@50-95收敛曲线对比图

下载: 全尺寸图片幻灯片

图 9 Jetson Orin Nano边缘部署实验环境

下载: 全尺寸图片幻灯片

表 1 不同上采样的检测性能对比

上采样方式	P(%)	R(%)	mAP50(%)	mAP50-95(%)	Params/M	GFLOPs
Nearest(基线)	95.5	87.6	90.0	59.4	2.38	5.2
Transposed Conv	96.68	88.47	92.03	59.68	2.7	5.6
CUS	97.4	89.2	92.1	60.8	2.38	5.2

下载: 导出CSV

表 2 基于DUT Anti-UAV数据集消融实验结果

模型	P(%)	R(%)	mAP@50(%)	mAP@50-95(%)	GFLOPs	Paras(M)	Size(MB)	FPS
A	95.5	87.6	90.0	59.4	5.2	2.38	5.4	245.98
A+D	97.6	88.9	92.2	60.6	5.0	2.30	5.3	233.6
A+C	97.4	89.2	92.1	60.8	5.2	2.38	5.4	215.03
A+P	97.0	88.7	91.9	60.2	5.2	2.36	5.4	254.28
A+D+C	97.4	89.9	92.2	61.4	5.0	2.30	5.3	204.7
A+D+C+P	98.7	90.0	92.4	61.7	5.0	2.29	5.2	210.71

下载: 导出CSV

表 3 DPC3k2内部消融实验结果

模型	P(%)	R(%)	mAP@50(%)	mAP@50-95(%)	GFLOPs	Paras(M)	Size(MB)	FPS
Baseline	95.5	87.6	90.0	59.4	5.2	2.38	5.4	245.98
+DSConv	96.67	89.06	91.77	60.04	5.1	2.33	5.3	238.09
+PConv	96.88	89.04	91.81	60.1	4.9	2.18	5.1	248.29
+DPC3k2	97.6	88.9	92.2	60.6	5.0	2.30	5.3	233.6

下载: 导出CSV

表 4 基于TIB-Net 数据集消融实验结果

模型	P(%)	R(%)	mAP@50(%)	mAP@50-95(%)	GFLOPs	Paras(M)	Size(MB)	FPS
A	88.8	87.6	89.8	36.4	5.2	2.38	5.4	249.67
A+D	90	90	92.1	38.9	5.0	2.29	5.2	235.21
A+C	88.9	87.8	90.1	36.5	5.2	2.38	5.4	215.28
A+P	88.8	89.8	91.1	37.1	5.2	2.36	5.4	256.23
A+D+C	89.3	88.7	91	37.9	5.0	2.30	5.3	204.88
A+D+C+P	90.7	88.9	91	39	5.0	2.29	5.2	212.07

下载: 导出CSV

表 5 YOLO 系列模型对比实验结果

模型	P(%)	R(%)	mAP@50(%)	mAP@50-95(%)	GFLOPs	Paras(M)	Size(MB)	FPS
YOLOv3-tiny	92.3	73.2	75.9	44.3	18.9	12.13	23.3	1038.62
YOLOv5n	96	84.3	87.4	56.3	7.1	2.5	5.3	451.7
YOLOv8n	95.7	84.7	87.2	56.5	8.1	3.0	6.3	501.51
YOLOv10n	95	86	88.5	56.5	6.5	2.27	5.8	297.44
YOLOv11n	95.9	86.1	88.7	57.2	6.3	2.58	5.5	360.53
YOLOv12n	95.7	85.9	88.1	56.7	6.3	2.56	5.6	213.45
YOLOv26n	95.5	87.6	90.0	59.4	5.2	2.38	5.4	245.98
Ours	98.7	90.0	92.4	61.7	5.0	2.29	5.2	210.71

下载: 导出CSV

表 6 主流模型对比实验结果

模型	P(%)	R(%)	mAP@50(%)	mAP@50-95(%)	GFLOPs	Paras(M)
AAP2-YOLO[11]	96.1	85.4	90.9	-	9.4	1.66
YOLO-ADD[12]	86.4	93	91.5	60.6	15.1	4.22
DCR-YOLO[13]	95	81.3	88.8	58.6	6.1	2.43
DRBD-YOLOv8[14]	90.9	77	85.7	53.5	5.7	1.57
YOLO11n-SMSH[15]	94.8	85.7	91.5	61.3	9.4	2.6
ET-YOLOv6[18]	-	-	90.6	58.8	7.9	3.2
DRF-YOLO[20]	93.9	80.3	86.9	54.8	-	-
TARG-YOLO[25]	-	-	89.98	61.14	75.64	19.7
Ours	98.7	90.0	92.4	61.7	5.0	2.29

下载: 导出CSV

表 7 VisDrone2019跨数据集泛化检测性能比较

模型	P(%)	R(%)	mAP@50(%)	mAP@50-95(%)
YOLOv26n	42.3	30.9	28.6	15.6
Ours	42.7	32.3	28.8	16.7

下载: 导出CSV

表 8 模型复杂度与推理速度比较

模型	GFLOPs	Paras(M)	Size(MB)	FPS
YOLOv26n	5.2	2.38	5.4	249.33
Ours	5.0	2.29	5.2	211.53

下载: 导出CSV

表 9 在Jetson Orin Nano平台上的边缘部署性能

模型	精度模式	mAP@50(%)	mAP@50-95(%)	FPS	内存/GB
YOLOv26n	FP32	90.98	60.14	95.05	3.07
DCP-YOLO	FP32	94.71	62.64	85.38	2.93
YOLOv26n	FP16	90.89	60.26	181.27	3.27
DCP-YOLO	FP16	94.59	62.7	116.02	3.00

下载: 导出CSV

参考文献(33)

[1]	聂伟, 戴琪霏, 杨小龙, 等. 基于多维信号特征的无人机探测识别方法[J]. 电子与信息学报, 2024, 46(3): 1089–1099. doi: 10.11999/JEIT230302. NIE Wei, DAI Qifei, YANG Xiaolong, et al. Unmanned aerial vehicle detection and recognition method based on multi-dimensional signal feature[J]. Journal of Electronics & Information Technology, 2024, 46(3): 1089–1099. doi: 10.11999/JEIT230302.
[2]	STUDIAWAN H, GRISPOS G, and CHOO K K R. Unmanned aerial vehicle (UAV) forensics: The good, the bad, and the unaddressed[J]. Computers & Security, 2023, 132: 103340. doi: 10.1016/j.cose.2023.103340.
[3]	WANG Bingshu, LI Qiang, MAO Qianchen, et al. A survey on vision-based anti unmanned aerial vehicles methods[J]. Drones, 2024, 8(9): 518. doi: 10.3390/drones8090518.
[4]	CHAUHAN D, KAGATHARA H, MEWADA H, et al. Nation’s defense: A comprehensive review of anti-drone systems and strategies[J]. IEEE Access, 2025, 13: 53476–53505. doi: 10.1109/ACCESS.2025.3550338.
[5]	GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]. 2014 IEEE Conference on Computer Vision and Pattern Recognition, Columbus, USA, 2014: 580–587. doi: 10.1109/CVPR.2014.81.
[6]	GIRSHICK R. Fast R-CNN[C]. 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, Chile, 2015: 1440–1448. doi: 10.1109/ICCV.2015.169.
[7]	HE Kaiming, GKIOXARI G, DOLLÁR P, et al. Mask R-CNN[C]. 2017 IEEE International Conference on Computer Vision (ICCV), Venice, Italy, 2017: 2980–2988. doi: 10.1109/ICCV.2017.322.
[8]	LIU Wei, ANGUELOV D, ERHAN D, et al. SSD: Single shot MultiBox detector[C]. Proceedings of 14th European Conference on Computer Vision – ECCV 2016, Amsterdam, The Netherlands, 2016: 21–37. doi: 10.1007/978-3-319-46448-0_2.
[9]	REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: Unified, real-time object detection[C]. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, USA, 2016: 779–788. doi: 10.1109/CVPR.2016.91.
[10]	SAPKOTA R, CHEPPALLY R H, SHARDA A, et al. YOLO26: Key architectural enhancements and performance benchmarking for real-time object detection[J]. arXiv preprint arXiv: 2509.25164, 2025. (查阅网上资料, 不确定本条文献类型是否正确, 请确认).
[11]	HU Xiang, ZHANG Xinding, HU Jianmin, et al. AAP2-YOLO: A model improvement about counter-UAV detection[C]. 2025 IEEE 6th International Conference on Computer, Big Data, Artificial Intelligence (ICCBD+AI), Xiamen, China, 2025: 1–6. doi: 10.1109/ICCBDAI66607.2025.11388615.
[12]	WANG Rijun, YANG Chunhui, TENG Xianglong, et al. YOLO-ADD: A high-precision detection model for small low-altitude drones in complex scenarios[J]. Applied Soft Computing, 2026, 199: 115347. doi: 10.1016/j.asoc.2026.115347.
[13]	DING Shaobo, ZHANG Maowei, LIU Datong, et al. DCR-YOLO: An enhanced anti-UAV detection method based on triple collaborative optimization strategy[J]. Aerospace Science and Technology, 2026, 168: 111319. doi: 10.1016/j.ast.2025.111319.
[14]	JIANG Panpan, YANG Xiaohua, WAN Yaping, et al. DRBD-YOLOv8: A lightweight and efficient anti-UAV detection model[J]. Sensors, 2024, 24(22): 7148. doi: 10.3390/s24227148.
[15]	YANG Jiyuan, GUO Tao, HAN Jinzu, et al. YOLO11n-SMSH: An improved UAV target detection model for YOLO11n[EB/OL]. https://www.preprints.org/manuscript/202509.0262, 2025.
[16]	吴林, 曹雯. YOLO-SCDI: 基于改进YOLOv8的车辆检测算法[J]. 电子与信息学报, 2025, 47(10): 3967–3978. doi: 10.11999/JEIT250238. WU Lin and CAO Wen. YOLO-SCDI: A vehicle detection algorithm based on an improved YOLOv8[J]. Journal of Electronics & Information Technology, 2025, 47(10): 3967–3978. doi: 10.11999/JEIT250238.
[17]	侯国家, 马佳琦, 王岳川, 等. UWF-YOLO: 冗余信息优化的轻量化水下目标检测[J]. 电子与信息学报, 2026, 48(7): 2876–2886. doi: 10.11999/JEIT251129. HOU Guojia, MA Jiaqi, WANG Yuechuan, et al. UWF-YOLO: A lightweight framework for underwater object detection via redundant information optimization[J]. Journal of Electronics & Information Technology, 2026, 48(7): 2876–2886. doi: 10.11999/JEIT251129.
[18]	贾亮, 林铭文, 戚丽瑾, 等. 面向无人机航拍图像的多尺度目标检测研究[J]. 半导体光电, 2024, 45(3): 501–507,514. doi: 10.16818/j.issn1001-5868.2024010903. JIA Liang, LIN Mingwen, QI Lijin, et al. Multiscale target detection for UAV aerial images[J]. Semiconductor Optoelectronics, 2024, 45(3): 501–507,514. doi: 10.16818/j.issn1001-5868.2024010903.
[19]	FENG Sheng, PU Xingchen, HU Yanting, et al. ET-YOLOv6: Tiny object detection algorithm on UAV targets based on star operation and cross-level fusion mechanism[EB/OL]. https://www.preprints.org/manuscript/202501.1348, 2025.
[20]	WANG Songwei, SHUAI Jianping, YANG Yuzhu, et al. DRF-YOLO model for small UAV detection through multi-scale residual enhancement and progressive feature fusion[J]. Applied System Innovation, 2025, 8(6): 179. doi: 10.3390/asi8060179.
[21]	方贤进, 姜雪凤, 徐留权, 等. CFS-YOLO: 一种利用粗细粒度搜索与焦点调制的早期火灾检测方法[J]. 电子与信息学报, 2025, 47(6): 1976–1991. doi: 10.11999/JEIT240928. FANG Xianjin, JIANG Xuefeng, XU Liuquan, et al. CFS-YOLO: An early fire detection method via coarse and fine grain search and focus modulation[J]. Journal of Electronics & Information Technology, 2025, 47(6): 1976–1991. doi: 10.11999/JEIT240928.
[22]	刘杰, 刘书豪, 田明, 等. 复杂环境下无人机航拍小目标检测算法[J]. 电子与信息学报, 2026, 48(4): 1763–1773. doi: 10.11999/JEIT251126. LIU Jie, LIU Shuhao, TIAN Ming, et al. Small object detection algorithm for UAV aerial images in complex environments[J]. Journal of Electronics & Information Technology, 2026, 48(4): 1763–1773. doi: 10.11999/JEIT251126.
[23]	肖振久, 赖思宇, 曲海成. 多层次精细化无人机图像目标检测[J]. 光电工程, 2025, 52(5): 240287. doi: 10.12086/oee.2025.240287. XIAO Zhenjiu, LAI Siyu, and QU Haicheng. Multi-level refined UAV image target detection[J]. Opto-Electronic Engineering, 2025, 52(5): 240287. doi: 10.12086/oee.2025.240287.
[24]	YANG Qiu, CHEN Chunmei, GONG Yuming, et al. RCF-YOLO: High precision small object visual detection algorithm for anti-UAV[J]. Measurement, 2026, 275: 121260. doi: 10.1016/j.measurement.2026.121260.
[25]	DU Yan, DAI Zifeng, WU Teng, et al. TARG-YOLO: An efficient small target detection framework for UAV[J]. Machine Vision and Applications, 2026, 37(2): 35. doi: 10.1007/s00138-026-01792-2.
[26]	贾小军, 唐彪, 骆淑云, 等. 基于频域感知特征融合的无人机航拍图像小目标检测算法[J]. 光电工程, 2026, 53(2): 250208. doi: 10.12086/oee.2026.250208. JIA Xiaojun, TANG Biao, LUO Shuyun, et al. A small target detection algorithm for UAV aerial images based on frequency domain perception feature fusion[J]. Opto-Electronic Engineering, 2026, 53(2): 250208. doi: 10.12086/oee.2026.250208.
[27]	HOWARD A G, ZHU Menglong, CHEN Bo, et al. MobileNets: Efficient convolutional neural networks for mobile vision applications[EB/OL]. https://arxiv.org/abs/1704.04861, 2017.
[28]	YANG Jiangnan, LIU Shuangli, WU Jingjun, et al. Pinwheel-shaped convolution and scale-based dynamic loss for infrared small target detection[C]. Proceedings of the 39th AAAI Conference on Artificial Intelligence, Philadelphia, USA, 2025: 9202–9210. doi: 10.1609/aaai.v39i9.32996.
[29]	HUANG Xuhong, LIU Shiqi, ZHANG Kai, et al. Reverse convolution and its applications to image restoration[C]. 2025 IEEE/CVF International Conference on Computer Vision (ICCV), Honolulu, USA, 2025: 10507–10516. doi: 10.1109/ICCV51701.2025.00978.
[30]	CAI Xinhao, LAI Qiuxia, WANG Yuwei, et al. Poly kernel inception network for remote sensing detection[C]. 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, USA, 2024: 27706–27716. doi: 10.1109/CVPR52733.2024.02617.
[31]	ZHAO Jie, ZHANG Jingshu, LI Dongdong, et al. Vision-based anti-UAV detection and tracking[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(12): 25323–25334. doi: 10.1109/TITS.2022.3177627.
[32]	SUN Han, YANG Jian, SHEN Jiaquan, et al. TIB-Net: Drone detection network with tiny iterative backbone[J]. IEEE Access, 2020, 8: 130697–130707. doi: 10.1109/ACCESS.2020.3009518.
[33]	ZHU Pengfei, WEN Longyin, DU Dawei, et al. Detection and tracking meet drones challenge[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 44(11): 7380–7399. doi: 10.1109/tpami.2021.3119563.