Generate Tie Points

Feature Description

Tie points are homologous image points that can construct stereo models or establish connections between adjacent models (images). Generating tie points helps improve accuracy during geometric correction to ensure spatial consistency of imagery.

Supported since SuperMap ImageX Pro 11i(2023) version.

Feature Entry

Imagery Tab->Geometric Correction Group->Generate Tie Points.

Parameter Description

  • Input Image Type: Select image type for generating tie points. Default is Panchromatic Image. Other options include Multispectral Image, Forward-looking Image, Rear-view Image, or Front View and Rear View Image.
  • Refer to Adjustment File: Use existing adjustment file information to make newly generated tie points approximate existing point accuracy. Manage multiple reference adjustment files using Add and Delete buttons on the toolbar.
  • Plane Accuracy: Determines whether preprocessing is required during tie point matching based on image planar accuracy.
    • Low: Indicates planar accuracy error greater than 40 pixels, requiring preprocessing.
    • High: Indicates planar accuracy error less than 15 pixels, no preprocessing needed.
    • Medium: Automatically estimates accuracy when uncertain about image precision.
  • Error Threshold: Set error threshold for outlier removal during image matching. Valid range [0,40], default 5 in px. Higher thresholds retain more points but increase error probability.
  • Point Distribution Mode: Select tie point distribution pattern. Options: Conventional and Uniform, default Conventional.
    • Conventional: Distributes points throughout overlap area with parameters:
      • Density: Set tie point density. Options: Sparse, Medium, Dense, default Medium.
      • Matching Method: Options: MOTIF, AFHORP, RIFT, SIFT, DEEPFT, default MOTIF. AFHORP and RIFT support multimodal data matching.
      • Maximum Points per Block: Maximum points retained per block during matching. Range [25,2048], default 256.
    • Uniform: Evenly distributes points in overlap area with parameters:
      • Seed Point Count: Set number of seed points per image. Range [64,6400], default 512. Increase for low-texture images to ensure sufficient matches.
      • Seed Point Search: Search methods: Grid Center and Corner Points, default Corner Points.
        • Corner Points: Selects distinctive points in region.
        • Grid Center: Uses grid center points with randomness.
      • Template Size: Interval between seed points. Range [1,256], default 40 in px. Larger templates increase reliability and processing time.
      • Search Radius: Matching search radius for seed points. Range [0,256], default 40 in px. Larger radii expand search scope and processing time.
    Figure: Regular Distribution Figure: Uniform Distribution
  • Semantic Culling of Non-ground Points: Disabled by default. When enabled, automatically removes tie points in cloud and building areas using AI semantic technology.
    • Cloud Area: Visible when semantic culling is enabled. Enabled by default to remove points in cloud areas via specified dataset. Disable to retain cloud area points.
    • Building Area: Visible when semantic culling is enabled. Enabled by default to automatically identify and remove building area points. Disable to retain building area points.

Related Topics

Manage Tie Points

Generate Ground Control Points

Manage Ground Control Points

Block Adjustment