Laser & Optoelectronics Progress, Volume. 62, Issue 16, 1601001(2025)

Underwater Image Object Detection Based on AWAF-YOLO Algorithm

Zhenghu Zhu1, Zhen Su1,2、*, and Wei Wang2
Author Affiliations
  • 1School of Automation, Jiangsu University of Science and Technology, Zhenjiang 212100, Jiangsu , China
  • 2Marine Equipment and Technology Institute, Jiangsu University of Science and Technology, Zhenjiang 212003, Jiangsu , China
  • show less
    Figures & Tables(14)
    Structure of YOLOv8 network
    Structure of AWAF-YOLO network
    Structure of ADown module
    Structure of WTConv module
    Structure of C2f-WTConv module
    Structure of SPPFAC module
    Schematic diagram of FSIoU
    Detection results of ablation experiments on the TrashCan1.0 dataset
    Detection results of ablation experiments on the Aquarium dataset
    Comparison of detection results for different models
    • Table 1. Comparative experimental results of loss function optimization

      View table

      Table 1. Comparative experimental results of loss function optimization

      ModelPRmAP@0.50mAP@0.50∶0.95
      YOLOv8n96.295.698.583.8
      YOLOv8n+Focaler-IoU96.195.998.584.0
      YOLOv8n+Shape-IoU95.796.498.684.0
      YOLOv8n+FSIoU95.996.498.684.3
      AWAF-YOLO96.796.998.886.4
    • Table 2. Ablation experimental results on the TrashCan1.0 dataset

      View table

      Table 2. Ablation experimental results on the TrashCan1.0 dataset

      ADownC2f-WTConvSPPFACFSIoUP /%R /%mAP@0.50 /%mAP@0.50∶0.95 /%Parameters /106FLOPS /109
      ××××96.295.698.583.83.08.2
      ×××95.796.298.584.02.67.3
      ×××95.995.998.583.62.67.2
      ×××95.797.098.784.83.88.9
      ×××95.896.498.684.33.08.2
      ××96.096.398.584.42.510.9
      ×96.596.898.786.23.411.6
      96.796.998.886.43.411.6
    • Table 3. Ablation experimental results on the Aquarium dataset

      View table

      Table 3. Ablation experimental results on the Aquarium dataset

      ADownC2f-WTConvSPPFACFSIoUP /%R /%mAP@0.50 /%

      mAP@

      0.50∶0.95 /%

      Parameters /106FLOPS /109
      ××××97.795.497.881.93.08.2
      ×××97.296.097.982.22.67.3
      ×××97.495.697.781.72.67.2
      ×××97.895.797.983.13.88.9
      ×××97.495.497.982.23.08.2
      ××97.495.897.882.12.510.9
      ×98.095.798.083.43.411.6
      97.896.198.183.53.411.6
    • Table 4. Comparative experimental results of different models

      View table

      Table 4. Comparative experimental results of different models

      ModelP /%R /%mAP@0.50 /%mAP@0.50∶0.95 /%Parameters /106FLOPS /109
      YOLOv3-tiny93.996.298.578.78.713.1
      YOLOv5s96.398.298.884.47.116.1
      YOLOv6s94.393.897.081.24.211.9
      YOLOv9s81.586.590.770.29.739.7
      YOLOv8n96.295.698.583.83.08.2
      AWAF-YOLO96.796.998.886.43.411.6
    Tools

    Get Citation

    Copy Citation Text

    Zhenghu Zhu, Zhen Su, Wei Wang. Underwater Image Object Detection Based on AWAF-YOLO Algorithm[J]. Laser & Optoelectronics Progress, 2025, 62(16): 1601001

    Download Citation

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

    Category: Atmospheric Optics and Oceanic Optics

    Received: Feb. 27, 2025

    Accepted: Mar. 19, 2025

    Published Online: Jul. 28, 2025

    The Author Email: Zhen Su (sz_just@126.com)

    DOI:10.3788/LOP250721

    CSTR:32186.14.LOP250721

    Topics