Acta Optica Sinica, Volume. 43, Issue 12, 1228012(2023)

Building Contour Optimization Method for Multi-Source Data

Xiang Hu, Jianhua Wu*, Ning Wei, and Haowen Tu
Author Affiliations
  • School of Geography and Environment, Jiangxi Normal University, Nanchang 330022, Jiangxi, China
  • show less
    Figures & Tables(19)
    General flow chart
    Comparison of simplified results. (a) Traditional D-P algorithm; (b) improved D-P algorithm; (c) stacking comparison
    Deviation of line fitting results caused by obtuse angle
    Feature edge regularization. (a) First type of vertical structure; (b) second type of vertical structure; (c) first type of parallel structure; (d) second type of parallel structure
    Feature angle regularization. (a) Type of removing obtuse angle; (b) type of removing acute angle
    Rule of right-angle. (a) One side tending to be perpendicular to main direction; (b) one side tending to be parallel to main direction; (c) one side neither tending to be perpendicular nor parallel to main direction
    Mesh Division in error circle. (a) Grids with n=2 in error circle; (b) grids with n=3 in error circle
    Comparison of contour position. (a) Contour before position refining; (b) contour after position refining; (c) result comparison
    Data sources and extracted building contours. (a) Historical raster map; (b) remote sensing image; (c) LiDAR point cloud data; (d) building contours extracted from historical raster map; (e) building contours extracted from remote sensing image; (f) building contours extracted from LiDAR point cloud data
    Contour comparison after superposition. (a) Superposition of optimization results and initial contours; (b) superposition of optimization results and historical raster map
    Comparison of optimized building contours (method A and proposed method). (a) Superposition of results obtained by method A and initial contours; (b) superposition of results obtained by proposed method and initial contours; (c) superposition of results obtained by method A and remote sensing image; (d) superposition of results obtained by proposed method and remote sensing image
    Comparison of optimized building contours (method B and proposed method). (a) Initial contour; (b) superposition of two methods; (c) superposition of proposed method and remote sensing image
    Optimized effect of building contours. (a) superposition of optimized result and initial contours; (b) superposition of optimized result and remote sensing image; (c) superposition of optimized result and street map
    Optimized effect of special shape building contours. (a)(e) Remote sensing image; (b)(f) initial contour; (c)(g) superposition of optimized result and initial contours; (d)(h) superposition of optimized result and remote sensing image
    • Table 1. Experimental arrangement and threshold selection

      View table

      Table 1. Experimental arrangement and threshold selection

      Experiment No.Data sourceνωαβγn
      1Historical raster map5.0d¯25d¯16030203
      2Remote sensing image8.0d¯48d¯17030103
      3LiDAR point cloud data2.1d¯8d¯16030203
    • Table 2. Accuracies of results of proposed method (experiment 1)

      View table

      Table 2. Accuracies of results of proposed method (experiment 1)

      IndexSimilarity of positionDegree of directional similarityDegree of size similarityDegree of shape similarity
      Value0.9660.9920.9690.986
    • Table 3. Result accuracy comparison of proposed method and method A

      View table

      Table 3. Result accuracy comparison of proposed method and method A

      MethodSimilarity of positionDegree of directional similarityDegree of size similarityDegree of shape similarity
      Method A0.9090.9480.8410.849
      Proposed method0.9120.9950.9850.975
    • Table 4. Accuracies of results of proposed method (Fig. 13)

      View table

      Table 4. Accuracies of results of proposed method (Fig. 13)

      IndexSimilarity of positionDegree of directional similarityDegree of size similarityDegree of shape similarity
      Value0.9540.9820.9620.965
    • Table 5. Time complexity

      View table

      Table 5. Time complexity

      Stage

      Simplification

      of contour

      Reconstruction of contourLocal regularization processing

      Right-angle

      of contour

      Refining position
      Time complexityOn2OnOnOnOn2
    Tools

    Get Citation

    Copy Citation Text

    Xiang Hu, Jianhua Wu, Ning Wei, Haowen Tu. Building Contour Optimization Method for Multi-Source Data[J]. Acta Optica Sinica, 2023, 43(12): 1228012

    Download Citation

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

    Category: Remote Sensing and Sensors

    Received: Nov. 4, 2022

    Accepted: Jan. 29, 2023

    Published Online: Jun. 20, 2023

    The Author Email: Wu Jianhua (wjhgis@126.com)

    DOI:10.3788/AOS221939

    Topics