• Laser & Optoelectronics Progress
  • Vol. 61, Issue 10, 1028001 (2024)
Xiuzai Zhang1、2、*, Tao Shen1, and Dai Xu1
Author Affiliations
  • 1School of Electronic and Information Engineering, Nanjing University of Information Science & Technology, Nanjing 210044, Jiangsu, China
  • 2Jiangsu Province Atmospheric Environment and Equipment Technology Collaborative Innovation Center, Nanjing University of Information Science & Technology, Nanjing 210044, Jiangsu, China
  • show less
    DOI: 10.3788/LOP231803 Cite this Article Set citation alerts
    Xiuzai Zhang, Tao Shen, Dai Xu. Remote-Sensing Image Object Detection Based on Improved YOLOv8 Algorithm[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1028001 Copy Citation Text show less
    YOLOv8 algorithm model structure
    Fig. 1. YOLOv8 algorithm model structure
    Improved YOLOv8 structure
    Fig. 2. Improved YOLOv8 structure
    Structure of GAM attention mechanism
    Fig. 3. Structure of GAM attention mechanism
    Channel attention submodule
    Fig. 4. Channel attention submodule
    Spatial attention submodule
    Fig. 5. Spatial attention submodule
    Residual attention module
    Fig. 6. Residual attention module
    Variable convolution process
    Fig. 7. Variable convolution process
    Deformable RoI pooling process
    Fig. 8. Deformable RoI pooling process
    Structure of C2f-DCN
    Fig. 9. Structure of C2f-DCN
    Detection effects of different models. (a)‒(c) Original images; (d)‒(f) detection effects of improved YOLOv3; (g)‒(i) detection effects of YOLOv5; (j)‒(i) detection effects of YOLOv8; (m)‒(o) detection effects of improved YOLOv8
    Fig. 10. Detection effects of different models. (a)‒(c) Original images; (d)‒(f) detection effects of improved YOLOv3; (g)‒(i) detection effects of YOLOv5; (j)‒(i) detection effects of YOLOv8; (m)‒(o) detection effects of improved YOLOv8
    MethodTypeP /%R /%mAP@0.5 /%FPS /(frame/s)
    1YOLOv875.566.169.898
    2YOLOv8+DWIoU75.865.970.295
    3YOLOv8+ReGAM74.966.770.392
    4YOLOv8+C2f-DCN76.166.571.081
    5YOLOv8+DWIoU+ReGAM75.866.670.989
    6YOLOv8+DWIoU+C2f-DCN76.166.771.580
    7YOLOv8+ReGAM+C2f-DCN75.966.671.378
    8YOLOV8+DWIoU+ReGAM+C2f-DCN76.266.872.177
    Table 1. Ablation experiments
    MethodFPS /(frame/s)mAP@0.5 /%
    Faster R-CNN842
    SSD3859.6
    YOLOv31868.2
    YOLOv41768.4
    YOLOv53869.4
    YOLOv89869.8
    Improved YOLOv87772.1
    Table 2. Performance comparison of different network models on DOTA dataset
    MethodFPS /(frame/s)mAP@0.5 /%
    Faster R-CNN1290.3
    SSD4689.2
    YOLOv33384.7
    YOLOv45787.8
    YOLOv56291.8
    YOLOv813793.4
    Improved YOLOv812394.6
    Table 3. Performance comparison of different network models on RSOD dataset
    Xiuzai Zhang, Tao Shen, Dai Xu. Remote-Sensing Image Object Detection Based on Improved YOLOv8 Algorithm[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1028001
    Download Citation