SpatialGrids provides 2D and 3D grid structures for working with point cloud data.

The current grids available are a 2D raster and a sparse voxel grid for working with 3D point cloud.

To create a 2D raster grid:

using SpatialGrids
points = rand(3, 100000) * 20.0
raster = rasterize_points(points, 0.1)

To create a sparse voxel grid with voxel side length of 10 metres for arbitrary points:

using SpatialGrids
points = rand(3, 100000) * 20.0
grid = SparseVoxelGrid(points, 10.0)

The created grid is an iteratable object that returns a Voxel at each index. Each voxel can be accessed directly with a for loop or all voxels can be collected into an array. Likewise, the returned Voxel is an iterable object that returns the point indices.

# Iterate through each voxel in grid
for voxel in grid
    # Get each point index in voxel
    for idx in voxel
        # Do stuff with points[:,idx]
    end
    # Or, you may want all point indices in a voxel
    all_point_indices = collect(voxel)
end