基于注意力和重构特征融合的轻量级煤矿安全帽检测方法

doi:10.3778/j.issn.1002-8331.2304-0421

摘要/Abstract

摘要： 针对当前的工人安全帽检测算法模型参数量较大、推理交互时间较长、错检和漏检率较高等问题，提出了一种基于注意力和重构特征融合的轻量级煤矿工人安全帽实时检测方法MS-YOLO。为了在不影响检测精度的前提下压缩模型的大小并提升检测的速度，采用轻量级网络MobileNeXt作为MS-YOLO算法的主干网络；重构了特征路径融合网络PANet，在网络中添加了新的尺度输入、ULSAM-4注意力模块和深度可分离卷积；为了加快模型收敛速度并提高预测框的回归精度，提出了一种新的损失函数CLIoU loss；该研究还建立了一个面向矿井场景的安全帽检测数据集以适用于其特殊的工况环境。通过在标准数据集和自建数据集上进行实验测试，结果表明，MS-YOLO模型不仅保持了较高的检测精度，还具有实时性好、模型轻量化的优点。

关键词: 目标检测, 轻量级网络, 煤矿安全帽检测, 损失函数, 注意力机制

Abstract: Aiming at the problems of large model parameters, long inference interaction time, high error detection and missing rate, a lightweight real-time detection method MS-YOLO for coal mine safety helmet based on attention and reconfiguration features fusion is proposed. Firstly, to compress the size of the model and improve the detection speed without affecting the detection accuracy, a lightweight network MobileNeXt is adopted as the backbone network of the MS-YOLO algorithm. Then, the feature path fusion network PANet is reconstructed, and a new scale input, ULSAM-4 attention module and depth-separable convolution are further added to the network. Moreover, to accelerate the convergence rate of the model and improve the regression accuracy of the prediction frame, a new loss function CLIoU loss is also proposed. Furthermore, a helmet detection dataset for mine scene is established to be suitable for its special working environment. Through the experimental test on the standard dataset and the self-built dataset, the results show that the proposed MS-YOLO not only maintains high detection accuracy, but also has the advantages of good real-time performance and lightweight model.

Key words: object detection, lightweight network, coal mine helmet detection, loss function, attention mechanism

董彦强, 程德强, 张云鹤, 寇旗旗, 张皓翔. 基于注意力和重构特征融合的轻量级煤矿安全帽检测方法[J]. 计算机工程与应用, 2024, 60(15): 297-306.

DONG Yanqiang, CHENG Deqiang, ZHANG Yunhe, KOU Qiqi, ZHANG Haoxiang. Lightweight Coal Mine Safety Helmet Detection Method Based on Attention and Reconfiguration Feature Fusion[J]. Computer Engineering and Applications, 2024, 60(15): 297-306.

参考文献

[1] 葛世荣, 郝尚清, 张世洪, 等. 我国智能化采煤技术现状及待突破关键技术[J]. 煤炭科学技术, 2020, 48(7): 28-46.
GE S R, HAO S Q, ZHANG S H, et al. Status of intelligent coal mining technology and potential key technologies in China[J]. Coal Science and Technology, 2020, 48(7): 28-46.
[2] 程德强, 徐进洋, 寇旗旗, 等. 融合残差信息轻量级网络的运煤皮带异物分类[J]. 煤炭学报, 2022, 47(3): 1361-1369.
CHENG D Q, XU J Y, KOU Q Q, et al. Lightweight network based on residual information for foreign body classification on coal conveyor belt[J]. Journal of China Coal Society, 2022, 47(3): 1361-1369.
[3] WANG X, NIU D, LUO P, et al. A safety helmet and protective clothing detection method based on Improved-YoloV3[C]//2020 Chinese Automation Congress (CAC), 2020: 5437-5441.
[4] REDMON J, FARHADI A. YOLOv3: an incremental improvement[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018: 101-104.
[5] LI Z, XIE W, ZHANG L, et al. Toward efficient safety helmet detection based on YoloV5 with hierarchical positive sample selection and box density filtering[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-14.
[6] 徐守坤, 王雅如, 顾玉宛, 等. 基于改进Faster RCNN的安全帽佩戴检测研究[J]. 计算机应用研究, 2020, 37(3): 901-905.
XU S K, WANG Y R, GU Y W, et al. Safety helmet wearing detection study based on improved Faster RCNN[J]. Application Research of Computers, 2020, 37(3): 901-905.
[7] SHEN J, XIONG X, LI Y, et al. Detecting safety helmet wearing on construction sites with bounding‐box regression and deep transfer learning[J]. Computer‐Aided Civil and Infrastructure Engineering, 2020, 36(2): 180-196.
[8] 谢斌红, 袁帅, 龚大立. 基于RDB-YOLOv4的煤矿井下有遮挡行人检测[J]. 计算机工程与应用, 2022, 58(5): 200-207.
XIE B H, YUAN S, GONG D L. Detection of blocked pedestrians based on RDB-YOLOv4 in coal mine[J]. Computer Engineering and Applications, 2022, 58(5): 200-207.
[9] LI N, GONG X. An image preprocessing model of coal and gangue in high dust and low light conditions based on the joint enhancement algorithm[J]. Computational Intelligence and Neuroscience, 2021: 2436486.
[10] BOCHKOVSKIY A, WANG C, LIAO H. Yolov4: optimal speed and accuracy of object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020: 9-12.
[11] WANG C, LIAO H, WU Y, et al. CSPNet: a new backbone that can enhance learning capability of CNN[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020: 390-391.
[12] LIU S, QI L, QIN H, et al. Path aggregation network for instance segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018: 8759-8768.
[13] ZHOU D Q, HOU Q B, CHEN Y P, et al. Rethinking bottleneck structure for efficient mobile network design[C]//Proceedings of the European Conference on Computer Vision(ECCV), 2020: 680-697.
[14] SANDLER M, HOWARD A, ZHU M, et al. MobileNetV2: inverted residuals and linear bottlenecks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018: 4510-4520.
[15] ZENG N, WU P, LI H, et al. A small-sized object detection oriented multi-scale feature fusion approach with application to defect detection[J]. IEEE Transactions on Instrumentation and Measurementvol, 2022, 71: 3507014.
[16] SAINI R, JHA N, DAS B, et al. ULSAM: ultra-lightweight subspace attention module for compact convolutional neural networks[C]//Proceedings of the IEEE Winter Conference on Applications of Computer Vision (WACV), 2020: 1616-1625.
[17] LU X C, JI J, XING Z, et al. Attention and feature fusion SSD for remote sensing object detection[J]. IEEE Transactions on Instrumentation and Measurement, 2021, 70: 1-9.
[18] 何丽, 张红, 房婉琳. 融合多尺度边界特征的显著实例分割[J]. 计算机科学与探索, 2022, 16(8): 1865-1876.
HE L, ZHANG H, FANG W L. Salient instance segmentation via multiscale boundary characteristic network[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(8): 1865-1876.
[19] 余震, 何留杰, 王振飞. 基于中智理论与方向α-均值的图像边缘检测算法[J]. 电子测量与仪器学报, 2020, 34(3): 43-50.
YU Z, HE L J, WANG Z F. Image edge detection based on intelligence theory and direction α-mean[J]. Journal of Electronic Measurement and Instrument, 2020, 34(3): 43-50.
[20] ZHENG Z, WANG P, LIU W, et al. Distance-IoU loss: faster and better learning for bounding box regression[C]//Thirty-Fourth AAAI Conference on Artificial Intelligence (AAAI), 2020: 12993-13000.
[21] LIU A J, LU J, ZHANG G Q, et al. Concept drift detection via equal intensity k-means space partitioning[J]. IEEE Transactions on Cybernetics, 2021, 51(6): 3198-3211.
[22] TAN M, PANG R, LE Q. EfficientDet: scalable and efficient object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2020: 10778-10787.
[23] GE Z, LIU S T, WANG F, et al. YOLOX: exceeding YOLO series in 2021[J]. arXiv:2107.08430, 2021.
[24] ZHANG Y, REN W, ZHANG Z, et al. Focal and efficient IOU loss for accurate bounding box regression[J]. Neurocomputing, 2022, 506: 146-157.
[25] 朱炳宇, 刘朕, 张景祥. 融合Grad-CAM和卷积神经网络的COVID-19检测算法[J]. 计算机科学与探索, 2022, 16(9): 2108-2120.
ZHU B Y, LIU Z, ZHANG J X. COVID-19 detection algorithm combining Grad-CAM and convolutional neural network[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(9): 2108-2120.