改进YOLOv8的轻量化无人机目标检测算法

doi:10.3778/j.issn.1002-8331.2310-0063

摘要/Abstract

摘要： 针对无人机目标检测算法计算复杂难以部署，且长尾分布的无人机数据导致检测精度较低的问题，提出了基于改进YOLOv8的轻量化无人机目标检测算法（PC-YOLOv8-n），可均衡网络检测精度与计算量，并对长尾分布数据有一定泛化能力。使用部分卷积层（PConv）替换YOLOv8中的3×3卷积层，对网络进行轻量化处理，解决网络冗余和计算量复杂的问题；融合双通道特征金字塔，增加自上而下的路径，将深层信息与浅层信息进行融合，同层引入轻量化注意力机制，提升网络的特征提取能力；采用均衡焦点损失（EFL）作为类别损失函数，通过均衡尾部类别在网络训练时的梯度权重，增加网络的类别检测能力。实验结果表明，PC-YOLOv8-n在VisDrone2019数据集中具有良好的表现，在mAP50精度上比原始YOLOv8-n算法提高了1.6个百分点，同时模型的参数和计算量分别降低为2.6×106和7.6 GFLOPs，检测速度达到77.2 FPS。

关键词: 无人机, YOLOv8, 长尾分布, 目标检测, 部分卷积

Abstract: Aiming at the problem that UAV target detection algorithms are computationally complex and difficult to deploy, and the long-tailed distribution of UAV data leads to low detection accuracy, a lightweight UAV target detection algorithm based on improved YOLOv8 (PC-YOLOv8-n) is proposed, which can balance the network detection accuracy and computation, and has some generalisation ability to long-tailed distribution of data. Using partial convolutional layers (PConv) to replace the 3×3 convolutional layers in YOLOv8, the network is lightweighted to solve the problems of network redundancy and computational complexity; it fuses dual-channel feature pyramids, increases top-down paths, fusion of deep and shallow information, and introduces a lightweight attention mechanism in the same layer to improve the feature extraction ability of the network; it uses the equilibrium focus loss (EFL) as the category loss function to increase the category detection ability of the network by equalising the gradient weights of the tail categories during network training. The experimental results show that PC-YOLOv8-n has good performance in the VisDrone2019 dataset, improving 1.6 percentage points in mAP50 accuracy over the original YOLOv8-n algorithm, while the parameters and computation of the model are reduced to 2.6×106 and 7.6 GFLOPs, respectively, and the detection speed reaches 77.2 FPS.

Key words: unmanned aerial vehicle (UAV), YOLOv8, long-tail distribution, object detection, partial convolution

胡峻峰, 李柏聪, 朱昊, 黄晓文. 改进YOLOv8的轻量化无人机目标检测算法[J]. 计算机工程与应用, 2024, 60(8): 182-191.

HU Junfeng, LI Baicong, ZHU Hao, HUANG Xiaowen. Improved YOLOv8 Lightweight UAV Target Detection Algorithm[J]. Computer Engineering and Applications, 2024, 60(8): 182-191.

参考文献

[1] 徐坚, 谢正光, 李洪均. 特征平衡的无人机航拍图像目标检测算法[J]. 计算机工程与应用, 2023, 59(6): 196-203.
XU J, XIE Z, LI H J. Feature-balanced UAV aerial image target detection algorithm[J].Computer Engineering and Applications, 2023, 59(6): 196-203.
[2] 马峻, 姚震, 徐翠锋, 等.基于改进PP-YOLO和Deep-SORT的多无人机实时跟踪算法[J].计算机应用, 2022, 42(9): 2885-2892.
MA J, YAO Z, XU C F, et al. Real-time multi-drone tracking algorithm based on improved PP-YOLO and Deep-SORT[J]. Journal of Computer Applications, 2022, 42(9): 2885-2892.
[3] CAO X, XU J, ZHANG R. Mobile edge computing for cellular-connected UAV: computation offloading and trajectory optimization[C]//Proceedings of the 2018 IEEE 19th International Workshop on Signal Processing Advances in Wireless Communications, 2018: 1-5.
[4] 张伟, 庄幸涛, 王雪力, 等.DS-YOLO: 一种部署在无人机终端上的小目标实时检测算法[J].南京邮电大学学报 (自然科学版), 2021, 41(1): 86-98.
ZHANG W, ZHUANG X T, WANG X L, et al. DS-YOLO: a real-time small object detection algorithm on UAVs[J].Journal of Nanjing University of Posts and Telecommunications (Natural Science Edition), 2021, 41(1): 86-98.
[5] 孙伟, 潘森, 黄恒.基于改进YOLOv4模型的无人机目标检测算法[J].传感技术学报, 2023, 36(3): 456-461.
SUN W, PAN S, HUANG H. UAV target detection algorithm based on improved YOLOv4 model[J]. Chinese Journal of Sensors and Actuators, 2023, 36(3): 456-461.
[6] 侯鑫, 曲国远, 魏大洲, 等.基于迭代稀疏训练的轻量化无人机目标检测算法[J].计算机研究与发展, 2022, 59(4): 882-893.
HOU X, QU G Y, WEI D Z, et al. A lightweight UAV object detection algorithm based on iterative sparse[J]. Journal of Computer Research and Development, 2022, 59(4): 882-893.
[7] 张红民, 庄旭, 郑敬添, 等. 优化YOLO网络的人体异常行为检测方法[J]. 计算机工程与应用, 2023, 59(7): 242-249.
ZHANG H M, ZHUANG X, ZHENG J T, et al.Optimizing human abnormal behavior detection method of YOLO network[J]. Computer Engineering and Applications, 2023, 59(7): 242-249.
[8] 陈范凯, 李士心. 改进Yolov5的无人机目标检测算法[J]. 计算机工程与应用, 2023, 59(18): 218-225.
CHEN F K, LI S X. UAV target detection algorithm with improved Yolov5[J]. Computer Engineering and Applications, 2023, 59(18): 218-225.
[9] YU W, YANG T, CHEN C. Towards resolving the challenge of long-tail distribution in UAV images for object detection[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021: 3258-3267.
[10] HUSSAIN M. YOLO-v1 to YOLO-v8, the rise of YOLO and its complementary nature toward digital manufacturing and industrial defect detection[J]. Machines, 2023, 11(7): 677.
[11] 王殿伟, 胡里晨, 房杰, 等.基于改进Double-Head RCNN的无人机航拍图像小目标检测算法[J/OL].北京航空航天大学学报: 1-10[2023-01-13].https://doi/10.13700/j.bh.1001-5965.2022.0591.
WANG D W, HU L, FANG J, et al. Small object detection algorithm based on improved Double-Head RCNN for UAV aerial images[J/OL]. Journal of Beijing University of Aeronautics and Astronautics: 1-10[2023-01-13].https://doi/10.13700/j.bh.1001-5965.2022.0591.
[12] 白宗宝, 张俊举, 高原, 等.基于注意力机制的航拍图像目标检测算法[J].激光与光电子学进展, 2023, 60(12): 312-322.
BAI Z B, ZHANG J J, GAO Y, et al. Attention mechanism-based object detection algorithm in aerial images[J]. Laser & Optoelectronics Progress, 2023, 60(12): 312-322.
[13] WANG F, WANG H, QIN Z, et al. UAV target detection algorithm based on improved YOLOv8[J].IEEE Access, 2023, 11: 116534-116544.
[14] LI Y, FAN Q, HUANG H, et al. A modified YOLOv8 detection network for UAV aerial image recognition[J]. Drones, 2023, 7(5): 304.
[15] WANG G, CHEN Y, AN P, et al. UAV-YOLOv8: a small-object-detection model based on improved YOLOv8 for UAV aerial photography scenarios[J]. Sensors, 2023, 23(16): 7190.
[16] TAN J, WANG C, LI B, et al. Equalization loss for long-tailed object recognition[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 662-671.
[17] SHU J, XIE Q, YI L, et al. Meta-Weight-Net: learning an explicit mapping for sample weighting[C]//Advances in Neural Information Processing Systems, 2019.
[18] WANG J, ZHANG W, ZANG Y, et al. Seesaw loss for long-tailed instance segmentation[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021: 9695-9704.
[19] CHEN J, KAO S, HE H, et al. Run, don’t walk: chasing higher FLOPS for faster neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023: 21-31.
[20] TAN M, PANG R, LE Q V. Efficientnet: scalable and efficient object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020: 781-790.
[21] ZAKI S Z, ZULKIFLEY M A, STOFA M M, et al. Classification of tomato leaf diseases using MobileNet v2[J]. IAES International Journal of Artificial Intelligence, 2020, 9(2): 290-296.
[22] MA N, ZHANG X, ZHENG H T, et al. ShuffleNet V2: practical guidelines for efficient CNN architecture design[C]//Proceedings of the European Conference on Computer Vision, 2018: 122-138.
[23] GONG Y, YU X, DING Y, et al. Effective fusion factor in FPN for tiny object detection[C]//Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021: 1160-1168.
[24] ZHANG T, ZHANG X, KE X. Quad-FPN: a novel quad feature pyramid network for SAR ship detection[J]. Remote Sensing, 2021, 13(14): 2771.
[25] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 7132-7141.
[26] DU D, ZHU P, WEN L, et al. VisDrone-DET2019: the vision meets drone object detection in image challenge results[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision Workshops, 2019: 213-226.
[27] LI B, YAO Y, TAN J, et al. Equalized focal loss for dense long-tailed object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022: 6990-6999.
[28] AKYON F C, ALTINUC S O, TEMIZELl A. Slicing aided hyper inference and fine-tuning for small object detection[C]//Proceedings of the 2022 IEEE International Conference on Image Processing, 2022: 966-970.
[29] BOTEV ZI, KROESE DP, RUBINSTEIN RY, et al. The cross-entropy method for optimization[J].Handbook of Statistics, 2013, 31: 35-59.
[30] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of the IEEE International Conference on Computer Vision, 2017: 2980-2988.
[31] ZHANG H, WANG Y, DAYOUB F, et al. Varifocalnet: an IoU-aware dense object detector[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021: 8514-8523.
[32] LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Proceedings of the European Conference on Computer Vision, 2016: 21-37.
[33] LONG X, DENG K, WANG G, et al. PP-YOLO: an effective and efficient implementation of object detector[J]. arXiv:2007.12099, 2020.
[34] GE Z, LIU S, WANG F, et al. YOLOX: exceeding YOLO series in 2021[J]. arXiv:2107.08430, 2021.
[35] JOCHER G, CHAURASIA A, STOKEN A, et al. Ultralytics/YOLOv5: v6.2-YOLOv5 classification models, Apple M1, reproducibility, ClearML and deci.ai integrations[EB/OL]. Zenodo (2022-08-17). https://zenodo.org/record/7002879.
[36] WANG C Y, BOCHKOVSKIY A, LIAO H Y M. YOLOv7: trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023: 7464-7475.