Tie Points

This pipeline extracts sparse tie points between two sensor images. It is used to improve geometric alignment during surface_modeling in the meta pipeline.

Allowed inputs

Tie points takes a single pair of sensor images as input, as shown in the input section of the documentation. Optionally, it can also take a rectification grid when called alone. The meta pipeline will make use of this optional input in order to not have to compute the rectification grid twice (once in surface_modeling and once in tie_points).

To specify rectification grids when using this pipeline alone, use the rectification_grid field in input:

input:
  sensors:
    image1: path/to/img_1.tif
    image2: path/to/img_2.tif

  rectification_grids:
    image1: path/to/grid_1.tif
    image2: path/to/grid_2.tif

{
    "input": {
        "sensors": {
            "image1": "path/to/img_1.tif",
            "image2": "path/to/img_2.tif"
        },
        "rectification_grids": {
            "image1": "path/to/grid_1.tif",
            "image2": "path/to/grid_2.tif"
        }
    }
}

Applications

The tie points pipeline uses these applications :

grid_generation

resampling

sparse_matching

Below is a schema summarizing how the Tie points pipeline works.

../../_images/tie_points_pipeline.drawio.png

Advanced Parameters

Name	Description	Type	Default value
save_intermediate_data	Save intermediate data for all applications inside this pipeline.	bool	False
geometry_plugin	Name of the geometry plugin to use and optional parameters (see geometry plugin)	str or dict	“SharelocGeometry”
resampling_tile_width	Width of the tiles used for resampling and SIFT computation, in pixels	int	5000
resampling_tile_height	Height of the tiles used for resampling and SIFT computation, in pixels	int	60

Below is an example configuration for this pipeline :

input:
  sensors:
    img1: path/to/img1.tif
    img2: path/to/img2.tif

  rectification_grids:
    img1: path/to/grid1.tif
    img2: path/to/grid2.tif

tie_points:
  applications:
    all:
      resampling:
        strip_height: 80
      sparse_matching:
        used_band: b2

  advanced:
    save_intermediate_data: true

{
    "input": {
        "sensors": {
            "img1": "path/to/img1.tif",
            "img2": "path/to/img2.tif"
        },
        "rectification_grids": {
            "img1": "path/to/grid1.tif",
            "img2": "path/to/grid2.tif"
        }
    },
    "tie_points": {
        "applications": {
            "all": {
                "resampling": {
                    "strip_height": 80
                },
                "sparse_matching": {
                    "used_band": "b2"
                }
            }
        },
        "advanced": {
            "save_intermediate_data": true
        }
    }
}