PCSED-YOLO：复杂环境下跨尺度多目标穿戴检测算法研究

doi:10.3778/j.issn.1002-8331.2505-0210

摘要/Abstract

摘要： 车间工人在作业期间若未按规定穿戴安全装备，不仅可能对其健康造成影响，还可能导致伤亡等安全事故。基于此，计算机视觉的穿戴检测技术是目前研究的热点领域。然而，由于生产车间内设备繁多，环境复杂且恶劣，生产过程中产生的大量粉尘会使穿戴目标变得模糊或部分被遮挡。此外，穿戴目标的尺寸分布范围宽，属于复杂环境下跨尺度多目标检测范畴。现有的算法在检测精度方面存在不足，特别是对口罩等小目标的误检和漏检率较高。为此，提出了一种基于YOLO模型的改进目标检测算法：PCSED-YOLO。在C3k2中融合了并行补丁感知模块，以增强小目标特征提取及多尺度目标检测能力；将交叉卷积注意力融合模块嵌入C2PSA，实现局部特征感知与全局上下文信息的互补，从而提升粉尘场景中的目标识别能力；引入空间到深度卷积替代原有的卷积层，通过重组空间维度信息至通道维度，实现无损下采样，提升小目标和低分辨率目标的检测性能；融合SEv2（squeeze-and-excitation network v2），创新改进空间金字塔池化层，增强模型对复杂场景的全局上下文把控能力，提升多类别、跨尺度目标的特征提取能力；在检测头引入动态卷积Dynamic-Conv，通过动态调整卷积核的大小和形状，提升跨尺度目标检测的精度；增加更高分辨率的P2检测层，提高小目标检测精度。制备了工人穿戴数据集，并进行了消融和对比实验。实验结果显示，PCSED-YOLO算法模型在处理小目标、中目标和大目标时均表现出色，与基准模型相比，mAP@0.5达到了0.946，提升了0.077；AP@0.5mask（小目标）达到了0.887，提升了0.236；AP@0.5no-helmet（中目标）提升了0.037至0.958；AP@0.5vest（大目标）提升了0.006至0.991；F1-Score和P-R曲线指标较基准模型也有明显改善。与几种先进的检测模型相比，PCSED-YOLO模型在制备的数据集上取得了最佳的检测性能，表明该模型具有较强的复杂环境跨尺度多目标检测能力和泛化能力，为复杂环境下跨尺度多目标穿戴检测提供了新的算法方案。

关键词: 安全穿戴检测, 小目标检测, 多尺度目标检测, 深度学习, YOLO

Abstract: Non-standard safety wear by workshop workers during operations can affect physical health and lead to safety accidents, including injuries or fatalities. Computer vision-based detection of personal protective equipment has become a research focus. Workshops have complex environments with numerous devices and harsh conditions, dust generated during production blurs or obscures targets, presenting challenges for multi-target detection across scales. Existing algorithms struggle with low accuracy and high false negative rates for small targets like masks. To address this, an improved YOLO-based detection algorithm, PCSED-YOLO, is proposed. The PPA module enhances feature extraction for small targets and multi-scale detection. The CAFM module in C2PSA achieves local and global feature complementarity, enhancing recognition in dusty environments. SPD-Conv replaces original convolutional layers for lossless down-sampling, improving detection of small, low-resolution targets. The SEv2 improvement of the spatial pyramid pooling layer enhances global context control and feature extraction for multi-class, cross-scale targets. Dynamic-Conv in the detection head adjusts convolution kernel size and shape for precise cross-scale detection. A higher resolution P2 detection layer improves small target accuracy. A dataset of workers’ protective wear is compiled for ablation and comparative experiments. Results show that PCSED-YOLO excels in handling targets of various sizes, achieving a mAP@0.5 of 0.946, a 0.077 improvement over the baseline. AP@0.5mask reaches 0.887, a 0.236 improvement, AP@0.5no-helmet is increased by 0.037 to 0.958, and AP@0.5vest is improved by 0.006 to 0.991. F1-Score and P-R curve metrics show significant enhancement. Compared to advanced models, PCSED-YOLO achieves the best accuracy on the dataset, demonstrating strong capabilities in complex environments and providing a new algorithm for multi-target detection of protective wear.

Key words: safety wear detection, small object detection, multi-scale object detection, deep learning, YOLO

薛光辉, 闫朝阳, 吴冕. PCSED-YOLO：复杂环境下跨尺度多目标穿戴检测算法研究[J]. 计算机工程与应用, 2026, 62(5): 88-105.

XUE Guanghui, YAN Zhaoyang, WU Mian. PCSED-YOLO： Study on Cross-Scale Multi-Object Wearable Detection Algorithm in Complex Environments[J]. Computer Engineering and Applications, 2026, 62(5): 88-105.

参考文献

[1] 冯爽，王万齐，杨文，等.基于改进RT-DETR的铁路施工场景下人员安全穿戴检测[J].铁道学报，2025，47（2）：92-101.
FENG S，WANG W Q，YANG W，et al.Safety wear detection for personnel in railway construction scenarios based on improved RT-DETR[J].Journal of the China Railway Society，2025，47（2）：92-101.
[2] 邱明明，刘超，胡正庭.基于Openpose-CenterNet的不停电作业人员安全防护用具穿戴智能检测研究[J].武汉大学学报（工学版），2024，57（6）：829-836.
QIU M M，LIU C，HU Z T.Research on intelligent detection of safety protective equipment worn by live workers of overhaul without power interruption based on Openpose-CenterNet[J].Engineering Journal of Wuhan University，2024，57（6）：829-836.
[3] 谢明志，樊丁萌，蒋志鹏，等.基于计算机视觉的混凝土结构裂缝检测研究现状与展望[J/OL].西南交通大学学报（2024-09-14）[2025-03-05].https：//link.cnki.net/urlid/51.1277.u.20240913.1853.002.
XIE M Z，FAN D M，JIANG Z P，et al.Research status and prospects of computer vision-based concrete structure crack detection[J/OL].Journal of Southwest Jiaotong University（2024-09-14）[2025-03-05].https：//link.cnki.net/urlid/51.1277.u.20240913.1853.002.
[4] 张金雷，杨健，刘晓冰，等.基于计算机视觉的轨道交通站内火灾检测与定位[J].交通运输系统工程与信息，2024，24（3）：53-63.
ZHANG J L，YANG J，LIU X B，et al.Computer vision-based fire detection and localization inside urban rail transit stations[J].Journal of Transportation Systems Engineering and Information Technology，2024，24（3）：53-63.
[5] 李华，吴立舟，薛曦澄，等.基于计算机视觉的高处临边作业安全巡检[J].中国安全科学学报，2023，33（9）：69-75.
LI H，WU L Z，XUE X C，et al.Computer vision based safety inspection of high abutting edges[J].China Safety Science Journal，2023，33（9）：69-75.
[6] 李春亚，王建华，尹土兵，等.个人防护装备穿戴安全检测技术的进展与应用[J].武汉理工大学学报，2024，46（10）：149-155.
LI C Y，WANG J H，YIN T B，et al.Advancements and applications of wearable safety detection technology for personal protective equipment[J].Journal of Wuhan University of Technology，2024，46（10）：149-155.
[7] 何嘉彬，李雷孝，林浩，等.面向计算机视觉的吸烟检测方法研究综述[J].计算机工程与应用，2024，60（1）：40-56.
HE J B，LI L X，LIN H，et al.Review of smoking detection methods for computer vision[J].Computer Engineering and Applications，2024，60（1）：40-56.
[8] 王伟，何姣，石强.复杂背景下目标识别算法分析与改进[J].智能计算机与应用，2020，10（4）：253-257.
WANG W，HE J，SHI Q.Analysis and improvement of target recognition algorithm in complex background[J].Intelligent Computer and Applications，2020，10（4）：253-257.
[9] 陈科圻，朱志亮，邓小明，等.多尺度目标检测的深度学习研究综述[J].软件学报，2021，32（4）：1201-1227.
CHEN K Q，ZHU Z L，DENG X M，et al.Deep learning for multi-scale object detection：a survey[J].Journal of Software，2021，32（4）：1201-1227.
[10] REN S Q，HE K M，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Mac-hine Intelligence，2017，39（6）：1137-1149.
[11] HE K M，GKIOXARI G，DOLLáR P，et al.Mask R-CNN[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2020，42（2）：386-397.
[12] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：unified，real-time object detection[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition.Piscataway：IEEE，2016：779-788.
[13] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot MultiBox detector[C]//Proceedings of the 14th European Conference on Computer Vision.Cham：Springer，2016：21-37.
[14] LIN T Y，GOYAL P，GIRSHICK R，et al.Focal loss for dense object detection[C]//Proceedings of the 2017 IEEE International Conference on Computer Vision.Piscataway：IEEE，2017：2999-3007.
[15] 王泉，叶广飞，陈祺东.YOLO-SWR：无人机视角下轻量级交通车辆检测算法[J].计算机工程与应用，2025，61（14）：112-122.
WANG Q，YE G F，CHEN Q D.YOLO-SWR：lightweight traffic vehicle detection algorithm from UAV perspective[J].Computer Engineering and Applications，2025，61（14）：112-122.
[16] 朱恩文，梁曌，肖进文，等.基于MDS-YOLO模型的小目标检测问题研究[J].湖南大学学报（自然科学版），2024，51（12）：78-86.
ZHU E W，LIANG Z，XIAO J W，et al.Research of small object detection problem based on MDS-YOLO model[J].Journal of Hunan University （Natural Sciences），2024，51（12）：78-86.
[17] 秦乐，谭泽富，雷国平，等.EMF-YOLO：轻量化多尺度特征提取路面缺陷检测算法[J].计算机工程与应用，2025，61（14）：101-111.
QIN L，TAN Z F，LEI G P，et al.EMF-YOLO：lightweight multi-scale feature extraction algorithm for road surface defect detection[J].Computer Engineering and Applications，2025，61（14）：101-111.
[18] 嵇晓峰.复杂场景下的安全穿戴检测研究与应用[D].东营：中国石油大学（华东），2022.
JI X F.Research and application of safe wear detection in complex scenes[D].Dongying：China University of Petroleum （Huadong），2022.
[19] XU S B，ZHENG S C，XU W H，et al.HCF-Net：hierarchical context fusion network for infrared small object detection[C]//Proceedings of the 2024 IEEE International Conference on Multimedia and Expo，2024.
[20] HU S，GAO F，ZHOU X W，et al.Hybrid convolutional and attention network for hyperspectral image denoising[J].IEEE Geoscience and Remote Sensing Letters，2024，21：5504005.
[21] SUNKARA R，LUO T.No more strided convolutions or pooling：a new CNN building block for low-resolution images and small objects[C]//Proceedings of the European Conference on Machine Learning and Knowledge Discovery in Databases.Cham：Springer，2023：443-459.
[22] NARAYANAN M.SENetv2：aggregated dense layer for channelwise and global representations[EB/OL].[2025-03-09].https：//arxiv.org/abs/2311.10807.
[23] WANG A，CHEN H，LIU L H，et al.YOLOv10：real-time end-to-end object detection[EB/OL].[2025-03-09].https：//arxiv.org/abs/2405.14458.
[24] HAN K，WANG Y H，GUO J Y，et al.ParameterNet：parameters are all you need for large-scale visual pretraining of mobile networks[C]//Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway：IEEE，2024：15751-15761.
[25] LAU K W，PO L M，REHMAN Y A U.Large separable kernel attention：rethinking the large kernel attention design in CNN[J].Expert Systems with Applications，2024，236：121352.
[26] SHAO H，ZENG Q S，HOU Q B，et al.MCANet：medical image segmentation with multi-scale cross-axis attention[J].Machine Intelligence Research，2025，22（3）：437-451.
[27] QIN D F，LEICHNER C，DELAKIS M，et al.MobileNetV4：universal models for the mobile ecosystem[C]//Proceedings of the 18th European Conference on Computer Vision.Cham：Springer，2025：78-96.
[28] SUN H，WEN Y，FENG H J，et al.Unsupervised bidirectional contrastive reconstruction and adaptive fine-grained channel attention networks for image dehazing[J].Neural Networks，2024，176：106314.
[29] AZAD R，NIGGEMEIER L，HüTTEMANN M，et al.Beyond self-attention：deformable large kernel attention for medical image segmentation[C]//Proceedings of the 2024 IEEE/CVF Winter Conference on Applications of Computer Vision.Piscataway：IEEE，2024：1276-1286.
[30] CAI X H，LAI Q X，WANG Y W，et al.Poly kernel inception network for remote sensing detection[C]//Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway：IEEE，2024：27706-27716.
[31] SI Y Z，XU H Y，ZHU X Z，et al.SCSA：exploring the synergistic effects between spatial and channel attention[EB/OL].[2025-03-12].https：//arxiv.org/abs/2407.05128.
[32] NI Z L，CHEN X H，ZHAI Y J，et al.Context-guided spatial feature reconstruction for efficient semantic segmentation[EB/OL].[2025-03-14].https：//arxiv.org/abs/2405.06228.
[33] YANG L X，ZHANG R Y，LI L D，et al.SimAM：a simple，parameter-free attention module for convolutional neural networks[C]//Proceedings of the International Conference on Machine Learning，2021.
[34] YAN C X，LI J F.An improved YOLOv5s algorithm and its application in object detection[J].Academic Journal of Computing & Information Science，2024，7（12）：36-43.
[35] YANG G Y，LEI J，ZHU Z K，et al.AFPN：asymptotic feature pyramid network for object detection[EB/OL].[2025-03-20].https：//arxiv.org/abs/2306.15988.
[36] ZHAO Y A，LV W Y，XU S L，et al.DETRs beat YOLOs on real-time object detection[C]//Proceedings of the 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway：IEEE，2024：16965-16974.
[37] TAN M X，LE Q V.EfficientNet：rethinking model scaling for convolutional neural networks[EB/OL].[2025-03-20].https：//arxiv.org/abs/1905.11946.
[38] YANG C，LI X K，ZHANG H L，et al.UAV small target det-ection in complex scenes based on improved YOLOv8s[C]//Proceedings of the 2024 39th Youth Academic Annual Conference of Chinese Association of Automation.Piscataway：IEEE，2024：1791-1798.
[39] FENG C J，ZHONG Y J，GAO Y，et al.TOOD：task-aligned one-stage object detection[C]//Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision.Piscataway：IEEE，2022：3490-3499.
[40] LIU Y H，LIU X C，ZHANG B.RetinaNet-vline：a flexible small target detection algorithm for efficient aggregation of information[J].Cluster Computing，2024，27（3）：2761-2773.
[41] LYU W Y，ZHAO Y，CHANG Q Y，et al.RT-DETRv2：imp-roved baseline with bag-of-freebies for real-time detection Transformer [EB/OL].[2025-03-25].https：//arxiv.org/abs/2407. 17140.