Remote Sensing Technology and Application, Volume. 40, Issue 1, 192(2025)

Classification Method for Rural Building Structures in Southeast Gansu Province based on Remote Sensing Images

Qinyao SUN1, Xiumei ZHONG1、*, Jinlian MA1, Yan WANG1, Xiaowei XU1, Songhan WU1, and Qian WANG1,3
Author Affiliations
  • 1Lanzhou Institute of Seismology,CEA,Lanzhou730000,China
  • 3Research Center for Conservation of Cultural Relics of Dunhuang,Dunhuang
  • show less
    Figures & Tables(9)
    Building extraction flowchart based on object-oriented technology
    Best scale of segmentation
    Classification effects of different segmentation scale
    Comparison of classification results between two methods
    • Table 1. GF-2 image remote sensing data information

      View table
      View in Article

      Table 1. GF-2 image remote sensing data information

      影像获取时间数据类型分辨率影像中心经纬度
      2021年8月7日GF-2全色0.8 m,多光谱3.2 m35.1°N,105.2°E
      2021年5月30日GF-2全色0.8 m,多光谱3.2 m35.2°N,105.2°E
    • Table 2. Primary feature factors

      View table
      View in Article

      Table 2. Primary feature factors

      特征特征因子特征特征因子
      纹理特征相关性光谱特征亮度
      方差平均值(4个)
      协同性标准差(4个)
      相异性形状特征长宽比
      对比度(2个)紧致度
      信息熵(2个)形状指数
      二阶矩(2个)自定义特征NDVI
      平均值(2个)SNDBI
      RVI
    • Table 3. Optimal feature extraction table for land features

      View table
      View in Article

      Table 3. Optimal feature extraction table for land features

      地物类别提取规则J-M距离E值区分类别Ce
      阴影SVI≤7501.973.88建筑物13.21
      1.943.53裸地
      1.872.96街道
      1.842.84植被
      植被NDVI≥0.651.993.02建筑物8.32
      1.972.79裸地
      1.942.51街道
      街道长宽比>0.35;亮度≥1 4001.871.73建筑物3.08
      1.511.35裸地
      裸地

      GLCD_信息熵<0.4;

      2 100≤亮度≤2 400

      1.991.63建筑物1.63
      土木建筑物GLCM_平均值>1251.951.54砖混建筑物1.54
      砖混建筑物GLCM_平均值<1251.951.54土木建筑物1.54
    • Table 4. Comparison of the accuracy of the classification of three kinds of building structures

      View table
      View in Article

      Table 4. Comparison of the accuracy of the classification of three kinds of building structures

      分类方法类型生产精度/%用户精度/%总体精度/%Kappa系数
      基于决策树的面向对象分类

      土木/砖木结构

      砖混结构

      93.6488.4682.420.70
      73.4992.34
      基于随机森林的面向对象分类土木/砖木结构94.8593.0286.820.76
      砖混结构77.3092.69
    • Table 5. Comparison of statistical data of building types with remote sensing interpretation results

      View table
      View in Article

      Table 5. Comparison of statistical data of building types with remote sensing interpretation results

      分类方法类型分类结果/栋目视解译结果/栋错分数量/栋错分率/%漏检数量/栋漏检率/%
      基于决策树面向对象分类

      土木/砖木结构

      砖混结构

      501

      327

      516

      348

      20

      14

      3.88

      4.02

      15

      21

      2.91

      5.46

      基于随机森林面向对象分类

      土木/砖木结构

      砖混结构

      497

      333

      516

      348

      12

      10

      2.33

      2.87

      19

      15

      3.68

      4.31

    Tools

    Get Citation

    Copy Citation Text

    Qinyao SUN, Xiumei ZHONG, Jinlian MA, Yan WANG, Xiaowei XU, Songhan WU, Qian WANG. Classification Method for Rural Building Structures in Southeast Gansu Province based on Remote Sensing Images[J]. Remote Sensing Technology and Application, 2025, 40(1): 192

    Download Citation

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

    Category:

    Received: Feb. 19, 2024

    Accepted: --

    Published Online: May. 22, 2025

    The Author Email: Xiumei ZHONG (xmzhong26@136.com)

    DOI:10.11873/j.issn.1004-0323.2025.1.0192

    Topics