filters.hexbin

filters.hexbin#

A common questions for users of point clouds is what the spatial extent of a point cloud collection is. Files generally provide only rectangular bounds, but often the points inside the files only fill up a small percentage of the area within the bounds.

Hexbin derived from input point buffer — Fig. 11 Hexbin output shows boundary of actual points in point buffer, not just rectangular extents.#

In addition to the original method of processing hexbins, density surfaces and boundaries can be processed using H3. This references hexbin products to a global grid of fixed hexagons at 16 resolutions, allowing joins and comparisons with other H3 datasets. When writing to a file with density, a unique H3Index is provided for each hexagon. Boundary smoothing is disabled for H3, and h3_resolution is used in place of edge_length.

The hexbin filter reads a point stream and writes out a metadata record that contains a boundary, expressed as a well-known text polygon. The filter counts the points in each hexagonal area to determine if that area should be included as part of the boundary. In order to write out the metadata record, the pdal pipeline command must be invoked using the “–pipeline-serialization” option:

Streamable Stage

This stage supports streaming operations

As an alternative to writing geometry to metadata, GDAL OGR can write to any OGR-compatible vector driver by specifying a filename with the density or boundary options. A valid driver that matches the file extension can be specified with ogrdriver; default is GeoJSON.

Example 1#

The following pipeline file and command produces an JSON output file containing the pipeline’s metadata, which includes the result of running the hexbin filter:

[
    "/Users/me/pdal/test/data/las/autzen_trim.las",
    {
        "type" : "filters.hexbin"
    }
]

$ pdal pipeline hexbin-pipeline.json --metadata hexbin-out.json

{
  "stages":
  {
    "filters.hexbin":
    {
      "area": 746772.7543,
      "avg_pt_per_sq_unit": 22.43269935,
      "avg_pt_spacing": 2.605540869,
      "boundary": "MULTIPOLYGON (((636274.38924399 848834.99817891, 637242.52219686 848834.99817891, 637274.79329529 849226.26445367, 637145.70890157 849338.05481789, 637242.52219686 849505.74036422, 636016.22045656 849505.74036422, 635983.94935813 849114.47408945, 636113.03375184 848890.89336102, 636274.38924399 848834.99817891)))",
      "boundary_json": { "type": "MultiPolygon", "coordinates": [ [ [ [ 636274.38924399, 848834.99817891 ], [ 637242.52219686, 848834.99817891 ], [ 637274.79329529, 849226.26445367 ], [ 637145.70890157, 849338.05481789 ], [ 637242.52219686, 849505.74036422 ], [ 636016.22045656, 849505.74036422 ], [ 635983.94935813, 849114.47408945 ], [ 636113.03375184, 848890.89336102 ], [ 636274.38924399, 848834.99817891 ] ] ] ] },
      "density": 0.1473004999,
      "edge_length": 0,
      "estimated_edge": 111.7903642,
      "hex_offsets": "MULTIPOINT (0 0, -32.2711 55.8952, 0 111.79, 64.5422 111.79, 96.8133 55.8952, 64.5422 0)",
      "sample_size": 5000,
      "threshold": 15
    }
},
...

Example 2#

As a convenience, the pdal info command will produce similar output:

$ pdal info --boundary /Users/me/test/data/las/autzen_trim.las

{
  "boundary":
  {
    "area": 746772.7543,
    "avg_pt_per_sq_unit": 22.43269935,
    "avg_pt_spacing": 2.605540869,
    "boundary": "MULTIPOLYGON (((636274.38924399 848834.99817891, 637242.52219686 848834.99817891, 637274.79329529 849226.26445367, 637145.70890157 849338.05481789, 637242.52219686 849505.74036422, 636016.22045656 849505.74036422, 635983.94935813 849114.47408945, 636113.03375184 848890.89336102, 636274.38924399 848834.99817891)))",
    "boundary_json": { "type": "MultiPolygon", "coordinates": [ [ [ [ 636274.38924399, 848834.99817891 ], [ 637242.52219686, 848834.99817891 ], [ 637274.79329529, 849226.26445367 ], [ 637145.70890157, 849338.05481789 ], [ 637242.52219686, 849505.74036422 ], [ 636016.22045656, 849505.74036422 ], [ 635983.94935813, 849114.47408945 ], [ 636113.03375184, 848890.89336102 ], [ 636274.38924399, 848834.99817891 ] ] ] ] },
    "density": 0.1473004999,
    "edge_length": 0,
    "estimated_edge": 111.7903642,
    "hex_offsets": "MULTIPOINT (0 0, -32.2711 55.8952, 0 111.79, 64.5422 111.79, 96.8133 55.8952, 64.5422 0)",
    "sample_size": 5000,
    "threshold": 15
  },
  "filename": "\/Users\/acbell\/pdal\/test\/data\/las\/autzen_trim.las",
  "pdal_version": "1.6.0 (git-version: 675afe)"
}

Options#

density: Output a density tessellation to the specified filename. ogrdriver must be compatible with the filename (default: GeoJSON FeatureCollection). If no file name is provided, nothing is written.
boundary: Output the grid’s boundary to the specified filename. ogrdriver must be compatible with the filename (default: GeoJSON FeatureCollection). If no file name is provided, nothing is written.
ogrdriver: GDAL OGR-compatible vector driver for writing with density or boundary. [Default: “GeoJSON”]
h3_grid: Create the hexbins using H3 hexagons. [Default: false]
h3_resolution: H3 resolution level the hexagons are created at (0, coarsest - 15, finest). Auto-calculates resolution if none is set.
edge_length: If not set, the hexbin filter will estimate a hex size based on a sample of the data. If set, hexbin will use the provided size in constructing the hexbins to test.
sample_size: How many points to sample when automatically calculating the edge size? Only applies if edge_length is not explicitly set. [Default: 5000]
threshold: Number of points that have to fall within a hexagon boundary before it is considered “in” the data set. [Default: 15]
precision: Minimum number of significant digits to use in writing out the well-known text of the boundary polygon. [Default: 8]
preserve_topology: Use GEOS SimplifyPreserveTopology instead of Simplify for polygon simplification with smooth option. [Default: true]
smooth: Use GEOS simplify operations to smooth boundary to a tolerance. Not compatible with H3 [Default: true]

where

An expression that limits points passed to a filter. Points that don’t pass the expression skip the stage but are available to subsequent stages in a pipeline. [Default: no filtering]

where_merge

A strategy for merging points skipped by a where option when running in standard mode. If true, the skipped points are added to the first point view returned by the skipped filter or if no views are returned, placed in their own view. If false, skipped points are placed in their own point view. If auto, skipped points are merged into the returned point view provided that only one point view is returned and it has the same point count as it did when the filter was run, otherwise the skipped points are placed in their own view.

[Default: auto]