面向轻量化网络的安全帽检测算法

doi:10.3778/j.issn.1002-8331.2105-0347

摘要/Abstract

摘要：

随着深度学习的发展，神经网络模型的体积越来越大，伴随而来的是参数量与计算量的增多，但实际安全帽检测环境下需要把网络模型部署在算力有限的移动端或嵌入式设备中，而这些设备无法支持复杂的计算量。针对这个问题，提出了一种适合部署在移动设备的轻量级目标检测网络HourGlass-YOLO（HG-YOLO）。以YOLOv5为基础模型，基于Inverted Resblock结构重构了新的主干特征提取网络HourGlass；并使用通道剪枝技术，对BatchNormalization（BN）层进行稀疏训练，将权值较小的通道进行删减，在保证精度的情况下，减少模型的参数；融合卷积层和BN层来加快在CPU上的推理速度。实验结果表明HG-YOLO在保证精度的情况下，将YOLOv5模型的体积压缩87%、浮点数减少86%、参数量降低89%，相比SSD在检测速度上快了8.2倍，更适合实际工业场景中的部署。

关键词: 深度学习, 目标检测, 轻量化模型, 剪枝

Abstract:

With the development of deep learning, the neural network model volume is becoming more larger, accompanied by the increase of parameters and computation. However, in helmet detection environment, the network model needs to be deployed in the mobile or embedded devices with limited computing power, and these devices can’t support complex computation. To solve this problem, this paper proposes a lightweight target detection network HourGlass YOLO（HG-YOLO） for mobile devices. Taking YOLOv5 as the basic model, a new backbone feature extraction network HourGlass is reconstructed based on Inverted Resblock structure. Channel pruning is used to sparsely train the BatchNormalization（BN） layer, and the channel with smaller weight is deleted to reduce the model parameters while ensuring the accuracy. Finally, the convolution and BN layer are integrated to speed up the reasoning speed on CPU. The experimental results show that HG-YOLO can compress the volume of YOLOv5 model by 87%, reduce the number of floating points by 86% and reduce the parameters number by 89%. Compared with SSD, HG-YOLO is 8.2 times faster in detection speed, which is more suitable for deployment in industrial scenes.

Key words: deep learning, target detection, lightweight model, pruning

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 Runxi, Alifu·Kuerban, GENG Liting. Safety Helmet Detection Algorithm for Lightweight Network[J]. Computer Engineering and Applications, 2021, 57(20): 263-270.

参考文献

[1] HSU Y L，TAI C Y，CHEN T C.Improving thermal properties of industrial safety helmets[J].International Journal of Industrial Ergonomics，2000，26（1）：109-117.
[2] 刘晓慧，叶西宁.肤色检测和Hu矩在安全帽识别中的应用[J].华东理工大学学报（自然科学版），2014，40（3）：365-370.
LIU Xiaohui，YE Xining.Skin color detection and Hu moment in helmet recognition research[J].Journal of East China University of Science and Technology（Natural Science Edition），2014，40（3）：365-370.
[3] 冯国臣，陈艳艳，陈宁，等.基于机器视觉的安全帽自动识别技术研究[J].机械设计与制造工程，2015，44（10）：39-42.
FENG Guochen，CHEN Yanyan，CHEN Ning，et al.Research on automatic identification technology of the safety helmet based on machine vision[J].Machine Design and Manufacturing Engineering，2015，44（10）：39-42.
[4] PARK M W，PALINGINIS E，BRILAKIS I.Detection of construction workers in video frames for automatic initialization of vision trackers[C]//Construction Research Congress 2012：Construction Challenges in a Flat World，2012：940-949.
[5] 李琪瑞.基于人体识别的安全帽视频检测系统研究与实现[D].成都：电子科技大学，2017.
LI Qirui.Research and implementation of helmet video detection system based on human body recognition[D].Chengdu：University of Electronic Science and Technology of China，2017.
[6] 许德刚，王露，李凡.深度学习的典型目标检测算法研究综述[J].计算机工程与应用，2021，57（8）：10-25.
XU Degang，WANG Lu，LI Fan.Review of typical object detection algorithms for deep learning[J].Computer Engineering and Applications，2021，57（8）：10-25.
[7] 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.
[8] HE K，ZHANG X，REN S，et al.Spatial pyramid pooling in deep convolutional networks for visual recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2015，37（9）：1904-1916.
[9] GIRSHICK R.Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision，2015：1440-1448.
[10] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[J].arXiv：1506.01497，2015.
[11] HE K，GKIOXARI G，DOLLáR P，et al.Mask R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision，2017：2961-2969.
[12] 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.
[13] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//European Conference on Computer Vision.Cham：Springer，2016：21-37.
[14] ZHOU X，ZHUO J，KRAHENBUHL P.Bottom-up object detection by grouping extreme and center points[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2019：850-859.
[15] 张明媛，曹志颖，赵雪峰，等.基于深度学习的建筑工人安全帽佩戴识别研究[J].安全与环境学报，2019，19（2）：535-541.
ZHANG Mingyuan，CAO Zhiying，ZHAO Xuefeng，et al.On the identification of the safe helmet wearing manners for the construction company workers based on the deep learning theory[J].Journal of Safety and Environment，2019，19（2）：535-541.
[16] 王兵，李文璟，唐欢.改进YOLO v3算法及其在安全帽检测中的应用[J].计算机工程与应用，2020，56（9）：33-40.
WANG Bing，LI Wenjing，TANG Huan.Improved YOLO v3 algorithm and its application in helmet detection[J].Computer Engineering and Applications，2020，56（9）：33-40.
[17] 施辉，陈先桥，杨英.改进YOLO v3的安全帽佩戴检测方法[J].计算机工程与应用，2019，55（11）：213-220.
SHI Hui，CHEN Xianqiao，YANG Ying.Safety helmet wearing detection method of improved YOLO v3[J].Computer Engineering and Applications，2019，55（11）：213-220.
[18] HINTON G，VINYALS O，DEAN J.Distilling the knowledge in a neural network[J].arXiv：1503.02531，2015.
[19] 高晗，田育龙，许封元，等.深度学习模型压缩与加速综述[J].软件学报，2021，32（1）：68-92.
GAO Han，TIAN Yulong，XU Fengyuan，et al.Summary of deep learning model compression and acceleration[J].Journal of Software，2021，32（1）：68-92.
[20] DENTON E，ZAREMBA W，BRUNA J，et al.Exploiting linear structure within convolutional networks for efficient evaluation[J].arXiv：1404.0736，2014.
[21] VANHOUCKE V，SENIOR A，MAO M Z.Improving the speed of neural networks on CPUs[C]//Proceedings of NIPS，2011.
[22] HAN S，POOL J，TRAN J，et al.Learning both weights and connections for efficient neural networks[J].arXiv：1506.02626，2015.
[23] REDMON J，FARHADI A.Yolov3：an incremental improvement[J].arXiv：1804.02767，2018.
[24] ARCOS-GARCIA A，ALVAREZ-GARCIA J A，SORIA-MORILLO L M.Evaluation of deep neural networks for traffic sign detection systems[J].Neurocomputing，2018，316：332-344.
[25] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2016：770-778.
[26] HOWARD A G，ZHU M，CHEN B，et al.Mobilenets：efficient convolutional neural networks for mobile vision applications[J].arXiv：1704.04861，2017.
[27] DETTMERS T，ZETTLEMOYER L.Sparse networks from scratch：faster training without losing performance[J].arXiv：1907.04840，2019.
[28] LI H，KADAV A，DURDANOVIC I，et al.Pruning filters for efficient convnets[J].arXiv：1608.08710，2016.
[29] GUILLEMOT M，HEUSELE C，KORICHI R，et al.Breaking batch normalization for better explainability of deep neural networks through layer-wise relevance propagation[J].arXiv：2002.11018，2020.
[30] DAQUAN Z，HOU Q，CHEN Y，et al.Rethinking bottleneck structure for efficient mobile network design[J].arXiv：2007.02269，2020.
[31] LIN T Y，DOLLáR P，GIRSHICK R，et al.Feature pyramid networks for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2017：2117-2125.