| 570 | 10 | 338 |
| 下载次数 | 被引频次 | 阅读次数 |
为了实现人群中跌倒行为的实时检测，预防踩踏事件的发生，针对跌倒行为检测实时性以及特征提取能力不足的问题，提出了一种改进YOLOv5s的跌倒行为检测算法。通过改进基本残差块，主干网络添加混合域注意力机制，颈部引入双向特征金字塔结构，以增强网络检测精度，同时保证运算量。结果表明，相比原始网络，所提算法检测准确率由94.1%提升到97.0%,精度值由91.2%提升到95.4%,且算法检测速度最快可达0.028 s,每秒检测图片帧数可达36.3,满足实时性要求。
Abstract：In order to realize the real-time detection of falling behavior in the crowd and prevent the occurrence of stampede events, an improved YOLOv5s fall behavior detection algorithm is proposed in view of the problems of real-time detection of falling behavior and insufficient feature extraction ability. The network detection accuracy is enhanced and the amount of computation is ensured by rational combination of improving the basic residual block, adding a mixed domain attention mechanism to the backbone network and introducing a bidirectional feature pyramid structure to the neck. The experimental results show that, compared with the original network, the detection accuracy of the proposed algorithm is improved from 94.1% to 97.0%, the accuracy value is increased from 91.2% to 95.4%, the optimized algorithm detection speed can reach 0.028 s, and the number of frames per second can reach 36.3, which meets the real-time requirements.
[1] 刘彦博，贾瑞生，徐志峰.基于尺度空间金字塔网络的人群计数算法[J].中国科技论文，2021,16(3):276-280.LIU Y B,JIA R S,XU Z F.Crowd counting algorithm based on scale space pyramid network [J].China Sciencepaper,2021,16(3):276-280.(in Chinese)
[2] JAEGER P F,KOHL S A A,BICKELHAUPT S,et al.Retina U-Net:embarrassingly simple exploitation of segmentation supervision for medical object detection [EB/OL].[2022-06-21].https://arxiv.org/abs/1811.08661.
[3] 栗俊杰，毛鹏军，方骞，等.基于K210的火点检测与喷头定向控制[J].科学技术与工程，2021,21(35):15136-15143.LI J J,MAO P J,FANG Q,et al.Fire detection and nozzle orientation control based on K210 [J].Science Technology and Engineering,2021,21(35):15136-15143.(in Chinese)
[4] 刘悦，张雷，辛山，等.融入时空注意力机制的深度学习网络视频动作分类[J].中国科技论文，2022,17(3):281-287.LIU Y,ZHANG L,XIN S,et al.Video action classification based on deep learning network with spatio-temporal attention mechanism [J].China Sciencepaper,2022,17(3):281-287.(in Chinese)
[5] 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 Machine Intelligence,2017,39(6):1137-1149.
[6] 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).New York:IEEE,2016:779-788.
[7] LIU W,ANGUELOV D,ERHAN D,et al.SSD:single shot multibox detector [C]// European Conference on Computer Vision.Cham:Springer,2016:21-37.
[8] LU N,WU Y D,FENG L,et al.Deep learning for fall detection:three-dimensional CNN combined with LSTM on video kinematic data [J].IEEE Journal of Biomedical and Health Informatics,2019,23(1):314-323.
[9] NOGAS J,KHAN S S,MIHAILIDIS A.DeepFall:non-invasive fall detection with deep spatio-temporal convolutional autoencoders[J].Journal of Healthcare Informatics Research,2020,4(1):50-70.
[10] 王兰燕.面向办案区倒地行为检测识别的深度学习算法研究[D].西安：西安电子科技大学，2021:53-60.WANG L Y.Research on deep learning algorithms for detection and recognition of falling behavior in case-handling areas [D].Xi’an:Xidian University,2021:53-60.(in Chinese)
[11] REDMON J,FARHADI A.YOLOv3:an incremental improvement [EB/OL].[2022-06-21].https://arxiv.org/abs/1804.02767.
[12] BOCHKOVSKIY A,WANG C Y,LIAO H Y.YOLOv4:optimal speed and accuracy of object detection [EB/OL].[2020-04-23].https://arxiv.org/abs/2004.10934v1.
[13] WANG C Y,LIAO H Y,WU Y H,et al.CSPNet:a new backbone that can enhance learning capability of CNN [C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).New York:IEEE,2020:1571-1580.
[14] ZHOU T T,ZHENG L,PENG Y P,et al.A survey of research on crowd abnormal behavior detection algorithm based on YOLO network [C]// 2022 2nd International Conference on Consumer Electronics and Computer Engineering (ICCECE).New York:IEEE,2022:783-786.
[15] 于硕，李慧，桂方俊，等.复杂场景下基于YOLOv5的口罩佩戴实时检测算法研究[J].计算机测量与控制，2021,29(12):188-194.YU S,LI H,GUI F J,et al.Research on real-time mask-wearing detection algorithm based on YOLOv5 in complex scenes [J].Computer Measurement & Control,2021,29(12):188-194.(in Chinese)
[16] LIN T Y,DOLLáR P,GIRSHICK R,et al.Feature pyramid networks for object detection [C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).New York:IEEE,2017:936-944.
[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.New York:IEEE,2018:8759-8768.
[18] REZATOFIGHI H,TSOI N,GWAK J,et al.Generalized intersection over union:a metric and a loss for bounding box regression [C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).New York:IEEE,2019:658-666.
[19] WOO S,PARK J,LEE J Y,et al.CBAM:convolutional block attention module [C]// European Conference on Computer Vision.Cham:Springer,2018:3-19.
[20] TAN M X,PANG R M,LE Q V.EfficientDet:scalable and efficient object detection [C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).New York:IEEE,2020:10778-10787.
[21] HOSSEIN R.Multicamera Human Action Video Data [DB/OL].[2022-06-21].http://dipersec.king.ac.uk/MuHAVi-MAS/.
[22] KEPSKI M.UR Fall Detection Dataset [DB/OL].[2022-06-21].http://fenix.ur.edu.pl/～mkepski/ds/uf.html.
[23] LI J F,WANG C,ZHU H,et al.CrowdPose:efficient crowded scenes pose estimation and a new benchmark [C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).New York:IEEE,2019:10855-10864.
[24] XIONG R B,YANG Y C,HE D,et al.On layer normalization in the transformer architecture [EB/OL].[2022-05-11].https://arxiv.org/abs/2002.04745v1.
基本信息:
中图分类号:TP391.41;TP183
引用信息:
[1]周彤彤,彭月平,郑璐,等.基于YOLOv5s的跌倒行为检测算法[J].中国科技论文,2023,18(07):786-792.
基金信息:
武警部队军事理论研究计划课题资助项目(WJJY21JL0205); 中国人民武装警察部队工程大学科研创新团队课题资助项目(KYTD201803)
2023-07-15
2023-07-15