目的 针对现代物流仓储场景中堆垛纸箱检测面临的复杂背景干扰、目标遮挡及小目标漏检等问题,提出一种改进的YOLOv8n目标检测方法YOLOv8n-CBF。方法 首先在骨干网络SPPF模块前嵌入CBAM注意力模块,抑制背景噪声;其次,采用双向特征金字塔网络（BiFPN）替换PANet结构,实现跨尺度特征加权双向融合;最后,增设160×160高分辨率小目标检测层Four,结合Distribution Focal Loss与Complete-IoU Loss提升定位精度。在公开的LSCD与OSCD数据集开展消融实验与对比实验。结果 消融实验表明,YOLOv8n-CBF算法在LSCD数据集上,平均精度均值（mAP@0.5）达到了84.6%,较YOLOv8n算法提升了3.3%。在OSCD数据集上,平均精度均值（mAP@0.5）达到了96.8%,较YOLOv8n算法提高了3.0%。与YOLO系列主流模型对比,改进模型检测精度最优,计算复杂度略有增加。结论 所提策略有效增强了模型特征提取与多尺度融合能力,显著提升了复杂场景检测精度与小目标感知能力,为自动化仓储管理提供了技术支撑。
Abstract
The work aims to propose an improved YOLOv8n object detection method named YOLOv8n-CBF to address the challenges of complex background interference, object occlusion, and missed detection of small objects in stacked carton detection within modern logistics warehousing scenarios. First, a Convolutional Block Attention Module (CBAM) was embedded before the SPPF module in the backbone network to suppress background noise. Second, a Bidirectional Feature Pyramid Network (BiFPN) was adopted to replace the PANet structure, achieving weighted bidirectional fusion of cross-scale features. Finally, a 160×160 high-resolution small-object detection layer (Four) was added, combined with Distribution Focal Loss and Complete-IoU Loss to improve localization accuracy. Ablation and comparative experiments were conducted on the publicly available LSCD and OSCD datasets. Ablation results demonstrated that the YOLOv8n-CBF algorithm achieved 84.6% mean average precision (mAP@0.5) on the LSCD dataset, representing an improvement of 3.3% over the YOLOv8n algorithm, and attained 96.8% mAP@0.5 on the OSCD dataset, surpassing YOLOv8n by 3.0%. Compared with mainstream YOLO series models, the proposed model achieves optimal detection accuracy with slightly increased computational complexity. The proposed strategies effectively enhance the model's feature extraction and multi-scale fusion capabilities, significantly improving detection accuracy and small-object perception in complex scenarios, thereby providing technical support for automated warehouse management.
关键词
纸箱堆叠 /
YOLOv8n /
注意力机制 /
多尺度特征增强 /
小目标检测
Key words
stacked carton /
YOLOv8n /
attention mechanism /
multi-scale feature enhancement /
small object detection
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 罗会兰, 陈鸿坤. 基于深度学习的目标检测研究综述[J]. 电子学报, 2020, 48(6): 1230-1239.
LUO H L, CHEN H K.Survey of Object Detection Based on Deep Learning[J]. Acta Electronica Sinica, 2020, 48(6): 1230-1239.
[2] DALAL N, TRIGGS B.Histograms of Oriented Gradients for Human Detection[C]//Proceedings of the International Conference on Image Processing.[s. l.]: IEEE, 2005.
[3] LOWE D G.Distinctive Image Features from Scale- Invariant Keypoints[J]. International Journal of Computer Vision, 2004, 60(2): 91-110.
[4] LIENHART R, MAYDT J.An Extended Set of Haar-Like Features for Rapid Object Detection[C]//Proceedings of the International Conference on Image Processing.[s. l.]: IEEE, 2002.
[5] 姜维, 张重生, 殷绪成. 基于深度学习的场景文字检测综述[J]. 电子学报, 2019, 47(5): 1152-1161.
JIANG W, ZHANG C S, YIN X C.Deep Learning Based Scene Text Detection: A Survey[J]. Acta Electronica Sinica, 2019, 47(5): 1152-1161.
[6] 林晨曦. 交互式堆垛纸箱机器视觉定位算法研究[D]. 武汉: 华中科技大学, 2021.
LIN C X.Research on Interactive Positioning Algorithm of Machine Vision for Stacked Cartons[D]. Wuhan: Huazhong University of Science and Technology, 2021.
[7] 陶磊, 李天剑, 胡欢. 基于改进Mask R-CNN的纸箱堆垛分割与定位方法[J]. 北京信息科技大学学报(自然科学版), 2020, 35(3): 85-88.
TAO L, LI T J, HU H.Carton Detection and Localization Method Based on the Improved Mask R-CNN[J]. Journal of Beijing Information Science & Technology University, 2020, 35(3): 85-88.
[8] 巩雪, 孙雪刚, 褚洋洋, 等. 基于改进Faster R-CNN的零食包装盒表面缺陷检测[J]. 包装工程, 2024, 45(23): 232-240.
GONG X, SUN X G, CHU Y Y, et al.Surface Defect Detection of Snack Packaging Box Based on Improved Faster R-CNN[J]. Packaging Engineering, 2024, 45(23): 232-240.
[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: IEEE, 2016.
[10] 陈翠琴, 范亚臣, 王林. 基于改进Mosaic数据增强和特征融合的Logo检测[J]. 计算机测量与控制, 2022, 30(10): 188-194.
CHEN C Q, FAN Y C, WANG L.Logo Detection Based on Improved Mosaic Data Enhancement and Feature Fusion[J]. Computer Measurement & Control, 2022, 30(10): 188-194.
[11] 李娇, 葛艳, 刘玉鹏. 基于改进YOLOv5的昏暗小目标交通标志识别[J]. 计算机系统应用, 2023, 32(5): 172-179.
LI J, GE Y, LIU Y P.Traffic Sign Recognition for Dim Small Targets Based on Improved YOLOv5[J]. Computer Systems & Applications, 2023, 32(5): 172-179.
[12] 杨明旭, 张俊宁, 张志强, 等. 基于改进YOLOv8的药片泡罩包装缺陷检测算法[J]. 包装工程, 2025, 46(1): 145-154.
YANG M X, ZHANG J N, ZHANG Z Q, et al.Defect Detection Algorithm of Pharmaceutical Blister Package Based on Improved YOLOv8[J]. Packaging Engineering, 2025, 46(1): 145-154.
[13] 牛馨予, 李世源, 赵剑道, 等. 结合改进YOLOv8的堆垛纸箱检测方法[J]. 制造业自动化, 2025, 47(5): 108-117.
NIU X Y, LI S Y, ZHAO J D, et al.Improved YOLOv8 Network Model for Stacked Carton Detection[J]. Manufacturing Automation, 2025, 47(5): 108-117.
[14] 贺鹏飞, 陈万新, 李伟, 等. MEP-Net: 基于YOLOv11n的铝制金属盖缺陷检测改进模型[J]. 包装工程, 2025, 46(21): 190-200.
HE P F, CHEN W X, LI W, et al.MEP-Net: An Improved Model for Defect Detection of Aluminum Metal Covers Based on YOLOv11n[J]. Packaging Engineering, 2025, 46(21): 190-200.
[15] WOO S, PARK J, LEE J Y, et al.CBAM: Convolutional Block Attention Module[C]//Computer Vision—ECCV 2018. Cham: Springer, 2018.
[16] LIN T Y, DOLLAR P, GIRSHICK R, et al.Feature Pyramid Networks for Object Detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu: IEEE, 2017.
[17] LIU S, QI L, QIN H F, et al.Path Aggregation Network for Instance Segmentation[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018.
[18] TAN M X, PANG R M, LE Q V.EfficientDet: Scalable and Efficient Object Detection[PP/OL]. V7. arXiv (2020-07-27). https://doi.org/10.48550/arXiv.1911.09070.
[19] 陈卫彪, 贾小军, 朱响斌, 等. 基于DSM-YOLO v5的无人机航拍图像目标检测[J]. 计算机工程与应用, 2023, 59(18): 226-233.
CHEN W B, JIA X J, ZHU X B, et al.Target Detection for UAV Image Based on DSM-YOLO V5[J]. Computer Engineering and Applications, 2023, 59(18): 226-233.
[20] YANG J R, WU S K, GOU L J, et al.SCD: A Stacked Carton Dataset for Detection and Segmentation[J]. Sensors, 2022, 22(10): 3617.
[21] LIN T Y, GOYAL P, GIRSHICK R, et al.Focal Loss for Dense Object Detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(2): 318-327.
[22] ZHANG L, SUN Z P, TAO H J, et al.Research on Mine-Personnel Helmet Detection Based on Multi- Strategy-Improved YOLOv11[J]. Sensors, 2024, 25(1): 170.
[23] TIAN Y J, YE Q X, DOERMANN D. YOLOv12: Attention-Centric Real-Time Object Detectors[PP/OL]. arXiv (2025-02-18). https://doi.org/10.48550/arXiv.2502.12524.
基金
国家自然科学基金（11901038）; 北京物资学院研究生科创项目（054260010635）
PDF(10809 KB)
渝公网安备 50010702501716号 渝ICP备15012534号-2
