计算机科学 ›› 2024, Vol. 51 ›› Issue (6A): 230900021-6.doi: 10.11896/jsjkx.230900021
李源鑫, 郭忠峰, 杨钧麟
LI Yuanxin, GUO Zhongfeng, YANG Junlin
摘要: 为提高现有集装箱的锁孔识别检测效率,减少算法参数量以及减小模型大小,提出了一种基于轻量化YOLOv5s的集装箱锁孔识别算法。该算法将YOLOv5s的Backbone主干特征提取网络部分更换为轻量级神经网络模型MobileNetV3,并对neck部分的特征融合结构进行进一步的优化,减少了模型的参数量和计算量,并提高了检测速度。引入注意力机制SimAM层,提高了检测的准确率和效率。使用不同的改进方法对模型进行重构后,在自建的集装箱锁孔数据集上进行训练和测试,并与改进的YOLOv5s进行对比实验。结果表明,改进后的模型大小仅为2.4MB,每幅图像的平均检测时间仅为5.1ms,平均检测精度达97.3%;与原始目标检测模型相比,该模型的大小减小了82.8%,检测速度提高了39%,在确保高检测精度的前提下展现出了较强的算法实时性。
中图分类号:
[1]CHANG Y M,ZHU X N,WANG L.Review on integratedscheduling of container terminal[J].Journal of Transportation Engineering,2019,12(1):136-146. [2]DONG M W,HUANG L F,GU Y.Review on green job scheduling optimization of container terminals[J].Journal of Chongqing Jiaotong University(Natural Science Edition),2022,41(11):7-14. [3]MI W J,ZHANG Z W,MI C.Study on container hanging holesrecognition algorithm based on machine vision[J].Chinese Journal of Construction Machinery,2016,14(5):399-402. [4]HAO Y S,LU B,LIU F,et al.Key point smoothing tracking ofcontainer keyhole based on centernet[J].Journal of Control Engineering,2021,28(11):2108-2113. [5]ZHANG J,DIAO Y F,CHENG W M,et al.Container keyholetracking and center positioning based on video stream[J].Journal of Computer Applications,2019,39(S2):216-220. [6]KUNIMITSU S,ASAMA H,KAWABATA K,et al.Development of crane vision for positioning container[J].Journal of Robotics and Mechatronics,2004(2):16. [7]MI C,WANG J,MI W,et al.Research on regional clustering and two-stage SVM method for container truck recognition[J].Discrete & Continuous Dynamical Systems,2018,12(4):1117-1133. [8]JOCHER G,STOKEN A,BOROVEC J,et al.Ultralytics/yolov5:v3.1-Bug fixes and performance improvements(Version v3.1)[EB/OL].(2020-10-29)[2021-01-19].https://gitlab.com/ultralytics/yolov5/-/releases/v3.1. [9]REDMON J,FARHADI A.YOLOv3:An incremental improvement[J].arXiv:1804.02767,2018. [10]BOCHKOVSKIY A,WANG C Y,LIAO H Y M.YOLOv4:Optimal speed and accuracy of object detection[J].arXiv:2004.10934,2020. [11]HOWARD A,SANDLER M,CHEN B,et al.Searching for mobilenetv3[C]//2019 IEEE/CVF International Conference on Computer Vision(ICCV).Seoul,Korea(South).New York:IEEE Press,2019:1314-1324. [12]SIFRE L,MALLAT,STEPHANE.Rigid-Motion scattering for texture classification[J].arXiv:1403.1687,2014. [13]HU J,SHEN L,ALBANIE S,et al.Squeeze-and-excitation networks[J].IEEE Transactions on Pattern Analysis And Machine Intelligence,2020,42(8):2011-2023. [14]GLOROT X,BORDES A,BENGIO Y.Deep sparse rectifierneural networks[J].Journal of Machine Learning Research,2011,15:315-323. [15]ZI B,YIN Z Q,LI Y C,et al.Fast mobile component location method for cable-driven parallel robots based on yolo model[J].Chinese Journal of Mechanical Engineering,2019,55(3):64-72. [16]YANG Y,DENG H.GC-YOLOv3:you only look once withglobal context block[J].Electronics,2020,9(8):1235. [17]WOO S,PARK J,LEE J Y,et al.CBAM:convolutional block attention module[C]//2018 15th European Conference on Computer Vision.Munich,Germany,2018:3-19. [18]YANG L,ZHANG R Y,LI L,et al.SimAM:A simple parameter-free attention module for convolutional neural networks[C]//International Conference on Machine Learning.PMLR.2021. |
|