计算机科学 ›› 2023, Vol. 50 ›› Issue (6A): 220500005-6.doi: 10.11896/jsjkx.220500005
谢溥轩1,2, 崔金荣1,2, 赵敏3
XIE Puxuan1,2, CUI Jinrong1,2, ZHAO Min3
摘要： 在电动车交通事故中,颅脑损伤致死是电动车骑行人员死亡的主要原因,而大多数电动车骑行人员很少佩戴头盔,因此通过将目标检测算法与道路摄像头结合来监管电动车骑行者头盔佩戴情况具有很强的现实意义。针对目前电动车头盔佩戴检测存在着目标相互遮挡漏检率较高、较小目标漏检率较高的问题,文中提出了一种改进的YOLOv5目标检测算法,用于实现对电动车头盔佩戴情况的检测。该方法首先在YOLOv5网络中添加通道注意力机制ECA-Net,使得模型能够更快地检测到目标特征,从而提高模型的检测性能;其次,使用Bi-FPN加权双向特征金字塔模块,实现对不同层级特征重要性的平衡,有利于改进小目标漏检问题;最后,使用Alpha-CIoU Loss的损失函数,提高模型定位的准确性。实验结果表明,该方法在3种场景下对电动车骑行人员头盔佩戴情况的检测精度均高于其他模型,平均精度达到95.8%,相比原网络检测精度有所提升,实现了电动车头盔佩戴情况的高精度检测。
中图分类号:
| [1]The Ministry of Public Security Traffic Mana-gement Bureau deployed to carry out the “one h-elmet and one belt” safety guarding action[J].Road Traffic Management,2020,35(5):9-10. [2]LIU C,WANG J T,WANG M Y.Application of SSD network with introduced visual mechanis-m in motorcycle helmet wearing detection[J].Journal of Electronic Measurement and Instrumentation,2021,35(3):144-151. [3]DING M D,LI Y X.Design and implementation of helmet wea-ring recognition system for cycl-ists based on YOLOv3[J].China New Communication,2021,23(13):44-45. [4]JIA J S,BAO Q J,TANG H M.Safety helmet wearing detection based on deformable component model[J].Computer Application Research,2016,33(3):953-956. [5]CHENG C C,HU S F.Research on intelligent detection of helmet wearing in intensive scenes[J].Intelligent Computers and Applications,2020,10(9):223-224. [6]REDMON J,DIVVALA S,GIRSHICK R,et al.You Only Look Once:Unified,Real-Time Objct Detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2016:779-788. [7]LIU W,ANGUELOV D,ERHAN D,et al.SSD:Single ShotMultiBox Detector[C]//Computer Vision-ECCV 2016:14th European Conference.Springer International Publishing,2016:21-37. [8]GIRSHICK R.Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision.2015:1440-1448. [9]TAN M,PANG R,LE Q V.EfficientDet:S-calable and Efficient Object Detection[C]//2020 IEEE/C-VF Conference on Computer Vision and Pattern Recognition(CVPR).IEEE,2020. [10]HE J,ERFANI S,MA X,et al.Alpha-IoU:A Family of Power Intersection over Union Losses for Bounding Box Regression[J].Advances in Neural Information Processing Systems,2021,34:20230-20242. [11]WANG Q,WU B,ZHU P,et al.ECA-Net:Efficie-nt Channel Attention for Deep Convolutional Neu-ral Networks[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR).IEEE,2020. [12]LIN T Y,DOLLAR P,GIRSHICKR,et al.Feature P-yramidNetworks for Object Detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2017:2117-2125. [13]LI H,XIONG P,AN J,et al.Pyramid Attention Network for Semantic Segmentation[J].arXiv:1805.10180,2018. [14]JIE H,LI S,GANG S,et al.Squeeze-and-Excitation Networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2018:7132-7141. [15]REN S,HE K,GIRSHICK R,et al.Faster R-CNN:Towards Real-Time Object Detection with Region Proposal Networks[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2017,39(6):1137-1149. |
|
||
首页