Journal of Applied Optics, Volume. 44, Issue 5, 1054(2023)

Infrared ship target detection algorithm based on improved YOLOX-S

Shuli LOU1... Yan WANG1,*, Jianqin GUO2 and Weifeng GONG3 |Show fewer author(s)
Author Affiliations
  • 1School of Physics and Electronic Information, Yantai University, Yantai 264005, China
  • 2Department of Electronics and Communication Engineering, Shandong College of Electronic Technology, Jinan 250200, China
  • 3State Key Laboratory of High-end Server and Storage Technology, Jinan 250101, China
  • show less
    Figures & Tables(8)
    YOLOX-S network structure model
    Improved YOLOX-S network structure model
    Traditional convolution structure
    Depthwise separable convolution structure
    ECANet channel attention mechanism
    Comparison of detection effect
    • Table 1. Comparison of different algorithms

      View table
      View in Article

      Table 1. Comparison of different algorithms

      ModelBackbonePrecision%AP%FPSParams
      Faster R-CNNResNet-500.980.974108.6 M
      SSDVGG0.950.924799.76 M
      YOLOv4MobileNetv20.940.925241.2 M
      YOLOv5-sCSPDarknet0.940.945427.76 M
      YOLOX-SCSPDarknet0.960.955034.21 M
      Improved YOLOX-SCSPDarknet0.980.985625.79 M
    • Table 2. Comparison of ablation experiment results

      View table
      View in Article

      Table 2. Comparison of ablation experiment results

      DSCECANetLoss_CIoUAP%FPSParams
      ---0.955034.21 M
      √---√---√0.940.970.9658495025.79 M34.21 M34.21 M
      -0.975625.79 M
      0.985625.79 M
    Tools

    Get Citation

    Copy Citation Text

    Shuli LOU, Yan WANG, Jianqin GUO, Weifeng GONG. Infrared ship target detection algorithm based on improved YOLOX-S[J]. Journal of Applied Optics, 2023, 44(5): 1054

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Research Articles

    Received: Sep. 21, 2022

    Accepted: --

    Published Online: Mar. 12, 2024

    The Author Email: WANG Yan (王岩)

    DOI:10.5768/JAO202344.0502006

    Topics