Deformation Detection

Feature Description

Deformation detection is used to identify stretching deformation areas such as mountains, roads, and buildings that occur after orthorectification or fusion. This method simplifies the workflow of manual inspection for deformation areas, significantly improving detection efficiency. Meanwhile, based on the detection result dataset, deformation areas can be quickly located and image repair can be implemented, enhancing the quality of the final result imagery.

SuperMap ImageX Pro Supported starting from version 11i(2024).

Function Entry

Imagery tab->Satellite data processing group->Deformation detection.

Parameter description

  • Detection Type: Provides three detection types: Terrain deformation, Road deformation, and Building deformation, all selected by default. You can select only one or two detection types according to project requirements.
  • Image to be checked: Supports input of orthorectified images, true color images, and fused images. When the image path contains both panchromatic and multispectral images, only the panchromatic image is loaded.
    • Add Image: Supports setting images for deformation detection through four methods: Add File, Add Folder, Add List, and Add Mosaic Dataset.
    • Add RPC file: If the image to be inspected is produced by SuperMap series products, there is no need to input the adjusted RPC file. If it is produced by other software, the adjusted RPC file must be imported.
  • Elevation Data: Sets DEM elevation data used to detect stretching deformation areas. It is recommended to use DEM data with a resolution of 30m or higher. Supports four formats: Tiff/GeoTIFF, Erdas Image, PCIDSK, and ArcInfoGrid. Among them, ArcInfoGrid format data can only be obtained by clicking the Add Folder button.
  • Detection range:
    • Range Data: Sets detection range area data. Supported formats include UDBX/UDB, GeoJson, ShapeFile.
  • Terrain deformation:
    • Deformation threshold: When the detected deformation degree is greater than the set threshold, it is marked as a stretching deformation area. The default value is 0.4, with a value range of (0.15,1.0). The smaller the value, the lower the stretching deformation degree in the identified results, but it increases the probability of false identification; the larger the value, the smaller the stretching deformation range in the identified image, with higher accuracy, but it increases the probability of missed identification.
  • Road deformation:
    • Road Line Data: Sets road line data. Supported formats include UDBX/UDB, GeoJson, ShapeFile.
    • Sampling point interval: First, sample the road line, and use the sampled result point dataset for deformation analysis. The default value is 15, in m.
    • Deformation threshold: When the detected deformation degree is greater than the set threshold, it is marked as a stretching deformation area. The default value is 0.2, with a value range of (0,1.0). The smaller the value, the lower the stretching deformation degree in the identified results, but it increases the probability of false identification; the larger the value, the smaller the stretching deformation range in the identified image, with higher accuracy, but it increases the probability of missed identification.
    • Road buffer radius: Sets the buffer radius for the detection result region dataset. The larger the value, the more obvious the detection result. The default value is 3, in m.
  • Building deformation:
    • Building Area: Sets the area data of the building region. Supported formats include UDBX/UDB, GeoJson, ShapeFile. If no building area is set, the AI model will automatically detect the building areas in the image, but the detection time will be longer.
    • Deformation threshold: When the detected deformation degree is greater than the set threshold, it is marked as a stretching deformation area. The default value is 0.5, with a value range of (0.4,1.0). The smaller the value, the lower the stretching deformation degree in the identified results, but it increases the probability of false identification; the larger the value, the smaller the stretching deformation range in the identified image, with higher accuracy, but it increases the probability of missed identification.
    • Pixel count threshold: When the pixel count of the stretching deformation area is greater than or equal to the set threshold, keep the detection result; otherwise, discard the area. The default value is 20.
    Detection result:
    • Datasource: Sets the datasource where the detection result data is stored. The detection result dataset names for terrain deformation, road deformation, and building deformation are: MountainDeformation, RoadDeformation, BuildingDeformation, and these names cannot be modified.

Related Topics

Plane Accuracy

Fine Plane Accuracy