vdbfusion 0.1.3

3D Volumetric Surface Reconstruction using the VDB data structure

NOTE: The support for this pip pacakge it is still under develpment and therefore might not be so stable.

VDBFusion

This is a small utility library that implement the VDBFusion algorithm, similar to TSDF-based reconstruction pipelines but using a different data-structure(VDB).

Installation

Take a seat and relax, you only need to:

pip install vdbfusion

If this step fails is most likely you have an old version of pip. To fix this, just pip install -U pip

Usage

from vdbfusion import VDBVolume

vdb_volume = VDBVolume(voxel_size=0.1, sdf_trunc=0.3, space_carving=False)

# You need to define your own Dataset, see snippet below
dataset = Dataset(...)

for scan, origin in dataset:
    vdb_volume.integrate(scan, origin)

Visualizing the output

Open3D(pip install open3d)

import open3d as o3d
# Extract a mesh from vdbfusion
vertices, triangles = vdb_volume.extract_triangle_mesh()

# Visualize the results
mesh = o3d.geometry.TriangleMesh(
    o3d.utility.Vector3dVector(vertices),
    o3d.utility.Vector3iVector(triangles),
)
mesh.compute_vertex_normals()
o3d.visualization.draw_geometries([mesh])

trimesh(pip install trimesh)

import trimesh

# Extract your map
vertices, triangles = vdb_volume.extract_triangle_mesh()

mesh = trimesh.Trimesh(vertices=vertices, faces=triangles)
mesh.show()

Dataset snippet

You Dataset class must only define a __len__ and __getitem__ methods, similar to torch.Dataset:

class Dataset:
    def __init__(self, *args, **kwargs):
        # Initialize your dataset here ..

    def __len__(self) -> int:
        return len(self.n_scans)

    def __getitem__(self, idx: int) -> Tuple[np.ndarray, np.ndarray]:
        # points: list of np.array points (N, 3)
        # origin: sensor origin on the global coordinate frame (np.ndarray(3))
        return points, origin