改进YOLOX的轻量级安全帽检测方法

doi:10.3778/j.issn.1002-8331.2204-0405

摘要/Abstract

摘要： 针对施工环境下安全帽数据集少，被检测物体目标小和现有检测模型参数量大导致的模型鲁棒性差，准确率低，训练时间长问题，提出了一种改进YOLOX网络的安全帽检测方法。使用在线困难样本挖掘（OHEM）寻找数据集中的困难样本，结合马赛克（Mosaic）方法对困难样本拼接来扩充训练集数量；在模型预测端（prediction）加入分支注意力模块，将网络输出分为两部分输入模块来提取空间层面和通道层面上关键信息；提出一种新的余弦退火算法，初始时加入预热（warm up），过程中逐段减小学习率曲线的振荡幅度，训练中减小模型的收敛时间。实验结果表明，与原方法相比，改进安全帽检测方法对安全帽检测的mAP、准确率、召回率分别提高了6.77、2.52、9.14个百分点，训练中使用CDWR余弦退火算法在同周期下损失值减少了0.5~1.0，与原算法相比训练收敛时间减少约50%。

关键词: 施工环境, 安全帽检测, YOLO网络, 注意力模块, 学习率

Abstract: Aiming at the problems of poor robustness, low accuracy and long training time caused by few helmet data sets, small detected object targets and large amount of existing detection model parameters in the construction environment, a helmet detection method based on improved YOLOX network is proposed. Firstly, online difficult sample mining（OHEM） is used to find the difficult samples in the dataset, and Mosaic method is used to splice the difficult samples to expand the number of training sets. Then the branch attention module is added to the prediction part of the model, and the network output is divided into two parts. The input module extracts the key information at the spatial level and channel level. Finally, a new cosine annealing algorithm is proposed, which adds warm up at the beginning, reduces the oscillation amplitude of the learning rate curve segment by segment in the process, and reduces the convergence time of the model in training. The experimental results show that compared with the original method, the improved helmet detection method improves the mAP, accuracy and recall of helmet detection by 6.77%, 2.52% and 9.14% respectively.Using the CDWR cosine annealing algorithm in training, the loss value is reduced by 0.5~1.0 in the same period, and the training convergence time is reduced by 50% compared with the original algorithm.

Key words: construction environment, helmet detection, YOLO network, attention module, learning rate

吕志轩, 魏霞, 马志钢. 改进YOLOX的轻量级安全帽检测方法[J]. 计算机工程与应用, 2023, 59(1): 61-71.

LV Zhixuan, WEI Xia, MA Zhigang. Lightweight Safety Helmet Detection Method of Improved YOLOX[J]. Computer Engineering and Applications, 2023, 59(1): 61-71.

参考文献

[1] DALAL N，TRIGGS B.Histograms of oriented gradients for human detection[C]//Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition，2005：886-893.
[2] VIOLA P，JONES M.Rapid object detection using a boosted cascade of simple features[C]//Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition，2001.
[3] FANG L J，LIN T W，WU J，et al.A DPM based approach to joint object detection and sub-category recognition[C]//Proceedings of the IEEE 7th Annual International Conference on CYBER Technology in Automation，2017：164-168.
[4] MARAYATR T，KUMHOM P.Motorcyclist’s helmet wearing detection using image processing[J].Advanced Materials Research，2014，931/932：588-592.
[5] GIRSHICK R，DONAHUE J，DARRELL T，et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2014：580-587.
[6] REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once：Unified，real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：779-788.
[7] LAW H，DENG J.CornerNet：Detecting objects as paired keypoints[C]//Proceedings of the European Conference on Computer Vision，2018.
[8] 王雨生，顾玉宛，庄丽华，等.复杂姿态下的安全帽佩戴检测方法研究[J].计算机工程与应用，2022，58（1）：190-196.
WANG Y S，GU Y W，ZHUANG L H，et al.Research on detection method of helmet wearing in complex posture[J].Computer Engineering and Applications，2022，58（1）：190-196.
[9] FU D S，GAO L，HU T，et al.Research on safety helmet detection algorithm of power workers based on improved YOLOv5[J].Journal of Physics（Conference Series），2022，2171：012006.
[10] 高明华，杨璨.基于改进卷积神经网络的交通目标检测方法[J].吉林大学学报（工学版），2022，52（6）：1353-1361.
GAO M H，YANG C.Traffic target detection method based on improved convolution neural network[J].Journal of Jilin University（Engineering and Technology Edition），2022，52（6）：1353-1361.
[11] 蒋润熙，阿里甫·库尔班，耿丽婷.面向轻量化网络的安全帽检测算法[J].计算机工程与应用，2021，57（20）：263-270.
JIANG R X，ALIFU·KUERBAN，GENG L T.Safety helmet detection algorithm for lightweight network[J].Computer Engineering and Applications，2021，57（20）：263-270.
[12] GE Z，LIU S T，WANG F，et al.YOLOX：Exceeding YOLO seriesin 2021[J].arXiv：2107.08430，2021.
[13] BOCHKOVSKIY A，WANG C Y，LIAO H Y M.YOLOv4：Optimal speed and accuracy of object detection[J].arXiv：2004.10934，2020.
[14] LEE H，KIM S.SSPNet：Learning spatiotemporal saliency prediction networks for visual tracking[J].Information Sciences，2021，575：399-416.
[15] SHRIVASTAVA A，GUPTA A，GIRSHICK R.Training region-based object detectors with online hard example mining[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：761-769.
[16] ELFWING S，UCHIBE E，DOYA K.Sigmoid-weighted linear units for neural network function approximation in reinforcement learning[J].Neural Networks，2018，107：3-11.
[17] CAO W H，FENG Z Y，ZHANG D Y，et al.Facial expression recognition via a CBAM embedded network[J].Procedia Computer Science，2020，174：463-477.
[18] JOSEPH J，MATHEW D K.Machine learning：Stochastic gradient descent in linear regression[J].Journal of Innovation and Social Science Research，2021，8（8）：176-178.
[19] POGGIO T，LIAO Q L，BANBURSKI A.Complexity control by gradient descent in deep networks[J].Nature Communications，2020，11：1027.
[20] 张娜，贺兴时.基于模拟退火的自适应正余弦算法[J].纺织高校基础科学学报，2021，34（1）：84-90.
ZHANG N，HE X S.A self-adaptive sine and cosine algorithm based on simulated annealing[J].Basic Sciences Journel Oftextile Universities，2021，34（1）：84-90.
[21] GOYAL P，P DOLLáR，GIRSHICK R，et al.Accurate，large minibatch SGD：Training ImageNet in 1 hour[DB/OL].（2017）[2022-04-01].https：//doc.paperpass.com/foreign/arXiv170602677.html.
[22] OTGONBOLD M E，GOCHOO M，ALNAJJAR F，et al.SHEL5K：An extended dataset and benchmarking for safety helmet detection[J].Sensors，2022，22（6）：2315.