could you help out converting it to .obj, please?

[AndreV84]

https://storage.googleapis.com/nvidia-dev/113620.ply At my end I either get the points lost or the color lost with conversion `[GCC 8.4.0] on linux Type "help", "copyright", "credits" or "license" for more information.

from pyntcloud import PyntCloud diamond = PyntCloud.from_file("cloud.ply") convex_hull_id = diamond.add_structure("convex_hull") convex_hull = diamond.structures[convex_hull_id] diamond.to_file("diamond_hull.obj", also_save=["mesh"])`

[AndreV84]

https://github.com/IntelRealSense/librealsense/issues/7800#issuecomment-729831606

[daavoo]

Hola @AndreV84 , I'm afraid that pyntcloud does not currently support reading/writing to/from .obj files with per-vertex color information.

As far as I remember when adding the .obj code, vertex coloring is not considered part of the format "standard" and it is only supported by a few applications (i.e. Meshlab).

Looking at how it appears to be implemented in Meshlab (https://gamedev.stackexchange.com/a/66270) it might be feasible to add support to vertex coloring for the .obj format in pyntcloud with minor modifications, will change the label to Feature Request. Contributions are allways wellcome.

[daavoo]

Hola again @AndreV84 I was feeling hacky so I think that you could use the following:

def write_obj(filename, points=None, mesh=None, save_vertex_colors=False):
    """
    Parameters
    ----------
    filename:   str
        The created file will be named with this
    points:     pd.DataFrame
    mesh:       pd.DataFrame
    save_vertex_colors: bool, optional
        Defaul False.
        If True, include per-vertex color info following:
        https://gamedev.stackexchange.com/a/66270

    Returns
    -------
    boolean
        True if no problems

    """
    if not filename.endswith('obj'):
        filename += '.obj'

    if points is not None:
        points = points.copy()
        columns_to_save = ["x", "y", "z"]
        if save_vertex_colors:
            columns_to_save.extend(["red", "green", "blue"])
            points["red"] /= 255.
            points["blue"] /= 255.
            points["green"] /= 255.
        points = points[columns_to_save]
        points.insert(loc=0, column="obj_v", value="v")
        points.to_csv(
            filename,
            sep=" ",
            index=False,
            header=False,
            mode='a',
            encoding='ascii')

    if mesh is not None:
        mesh = mesh.copy()
        mesh = mesh[["v1", "v2", "v3"]]
        mesh += 1  # index starts with 1 in obj file
        mesh.insert(loc=0, column="obj_f", value="f")
        mesh.to_csv(
            filename,
            sep=" ",
            index=False,
            header=False,
            mode='a',
            encoding='ascii')

    return True

cloud = PyntCloud.from_file("examples/data/ankylosaurus_mesh.ply")
write_obj("output.obj", points=cloud.points, mesh=cloud.mesh, save_vertex_colors=True)

[AndreV84]

Thank you very much! It seems I was able to export ply to obj with meshlab. I can visualize the obj in meshlab. Not sure if it wil get drawn in other visualizers though.

Could you elaborate on how to visualize from terminal with the python tool? I never got any window to pop up

[daavoo]

Thank you very much! It seems I was able to export ply to obj with meshlab. I can visualize the obj in meshlab. Not sure if it wil get drawn in other visualizers though.

Could you elaborate on how to visualize from terminal with the python tool? I never got any window to pop up

If you want to visualize directly from the terminal, you need to first install matplotlib (pip install matplotlib). And then:

from pyntcloud import PyntCloud
cloud = PyntCloud.from_file("examples/data/ankylosaurus_mesh.ply")
cloud.plot(backend="matplotlib")

[AndreV84]

Thank you very much!