• Spacecraft Recovery & Remote Sensing
  • Vol. 45, Issue 4, 139 (2024)
Shufeng HAO1, Yu GAO1, Ping LIU1,*, Yuang LI2..., Huadong ZHANG3, Hongjie REN1, Shuaijie TIAN1 and Wentao KOU4|Show fewer author(s)
Author Affiliations
  • 1College of Computer Science and Technology (College of Data Science), Taiyuan University of Technology, Taiyuan, 030024, China
  • 2College of Software, Taiyuan University of Technology, Taiyuan, 030024, China
  • 3College of Hydro Science and Engineering, Taiyuan University of Technology, Taiyuan, 030024, China
  • 4College of Mechanical and Vehicle Engineering, Taiyuan University of Technology, Taiyuan, 030024, China
  • show less
    DOI: 10.3969/j.issn.1009-8518.2024.04.014 Cite this Article
    Shufeng HAO, Yu GAO, Ping LIU, Yuang LI, Huadong ZHANG, Hongjie REN, Shuaijie TIAN, Wentao KOU. Flue Cured Tobacco Area Extraction of Remote Sensing Image by Integrating Lightweight ASPP and U-Net[J]. Spacecraft Recovery & Remote Sensing, 2024, 45(4): 139 Copy Citation Text show less
    [in Chinese]
    Fig. 1. [in Chinese]
    [in Chinese]
    Fig. 2. [in Chinese]
    [in Chinese]
    Fig. 3. [in Chinese]
    [in Chinese]
    Fig. 4. [in Chinese]
    [in Chinese]
    Fig. 5. [in Chinese]
    [in Chinese]
    Fig. 6. [in Chinese]
    [in Chinese]
    Fig. 7. [in Chinese]
    [in Chinese]
    Fig. 8. [in Chinese]
    [in Chinese]
    Fig. 9. [in Chinese]
    名称环境配置
    CPUAMD A10-7300,内存为8 GB
    图形处理器GPUTelas V100,内存为16 GB
    编程语言python3.5
    深度学习框架TensorFlow1.14.0+keras2.3.1
    CUDA版本9.0
    CUDNN版本7.6.5
    Table 1. The experimental setup
    算法优化器迭代次数/次学习率批次大小
    FCNAdam1000.000 116
    U-NetAdam1000.00132
    SegNetAdam1000.00132
    DeepLabv3+Adam1000.00132
    本文方法Adam1000.000132
    Table 2. The parameters setting
    算法PA/%MIoU/%参数量/MByte训练时间/h单图预测时间/ms
    U-Net92.983.195.00.5725
    FCN91.276.919.10.2212
    DeepLabv3+92.482.6209.41.0240
    SegNet92.081.044.10.3015
    本文方法93.784.195.20.5826
    Table 3. The comparison of different tobacco area segmentation methods
    算法PA/%MIoU/%参数量/MByte训练时间/h单图预测时间/ms
    U-Net92.983.195.00.5725
    方案一93.183.295.00.5725
    方案二93.684.195.50.5927
    本文方法93.784.195.20.5826
    Table 4. The results of ablation study
    Shufeng HAO, Yu GAO, Ping LIU, Yuang LI, Huadong ZHANG, Hongjie REN, Shuaijie TIAN, Wentao KOU. Flue Cured Tobacco Area Extraction of Remote Sensing Image by Integrating Lightweight ASPP and U-Net[J]. Spacecraft Recovery & Remote Sensing, 2024, 45(4): 139
    Download Citation