Journal of Geo-information Science, Volume. 22, Issue 10, 2051(2020)

Land Type Interpretation Authenticity Check of Vector Patch Supported by Deep Learning and Remote Sensing Image

Zihui GUO1 and Wei LIU1,2、*
Author Affiliations
  • 1School of Geographic Mapping and Urban Rural Planning, Jiangsu Normal University, Xuzhou 221116, China
  • 2State Key Laboratory of Resources and Environmental Information Systems, Institute of Geographic Sciencesand Natural Resources Research, Chinese Academy of Sciences, Beijing 100101, China
  • show less
    Figures & Tables(18)
    A sketch of the task of checking the authenticity of the land type interpretation of vector patches
    The technical route of the check of the authenticity of the land type interpretation of vector patches
    Research area overview and samples
    Purification process of self-labeled samples
    Inception module
    Structure diagram of the Inception_v3 model
    Effect of learning rate attenuation coefficient on model accuracy
    Inception_v3 learning rate dynamic curve
    Training accuracy change of Inception_v3 model
    Optimal grid schema in authenticity check of geographic interpretation
    Some suspicious spots
    Land use types with low characteristic differentiation
    The results of the check of the authenticity of the land type interpretation of vector patches in Dawu Town
    • Table 1. [in Chinese]

      View table
      View in Article

      Table 1. [in Chinese]

      类别训练集数量验证集数量测试集数量合计
      工业用地7971002001097
      林地7631082171088
      耕地90331214242912 676
      水体66594189948
      住宅用地22223176343173
      合计13 4801833366918 982
    • Table 2. Comparison of model effects used in scene classification

      View table
      View in Article

      Table 2. Comparison of model effects used in scene classification

      学习方法模型验证集精度测试集精度
      迁移学习VGG160.9160.902
      Inception_v30.9660.934
    • Table 3. Super parameter information

      View table
      View in Article

      Table 3. Super parameter information

      参数名称批处理值/(个/次)步数/初始/学习率学习率/衰减系数学习率衰减速度/(步/次)
      参数值100200 0000.100.9010 000
    • Table 4. Model classification results

      View table
      View in Article

      Table 4. Model classification results

      精确率(P)召回率(R)F1准确率(OA)
      工业用地0.9090.8750.8920.934
      林地0.9650.9250.945
      耕地0.9290.9360.933
      水体0.8990.9100.905
      住宅0.9810.9940.987
      均值0.9370.9280.932
    • Table 5. Effect of the check of the authenticity of the land type interpretation of vector patches

      View table
      View in Article

      Table 5. Effect of the check of the authenticity of the land type interpretation of vector patches

      准确率(A)精确率(P)召回率(R)
      0.7660.9250.817
    Tools

    Get Citation

    Copy Citation Text

    Zihui GUO, Wei LIU. Land Type Interpretation Authenticity Check of Vector Patch Supported by Deep Learning and Remote Sensing Image[J]. Journal of Geo-information Science, 2020, 22(10): 2051

    Download Citation

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

    Received: Jan. 1, 2020

    Accepted: --

    Published Online: Apr. 23, 2021

    The Author Email: Wei LIU (liuw@jsnu.edu.cn)

    DOI:10.12082/dqxxkx.2020.200001

    Topics