Open3D-PointNet2-Semantic3D icon indicating copy to clipboard operation
Open3D-PointNet2-Semantic3D copied to clipboard

Down sampling error

Open VMohinder opened this issue 3 years ago • 2 comments

Processing: bildstein_station1_xyz_intensity_rgb Num points: 29697591 Num points after 0-skip: 9476296 Traceback (most recent call last): File "downsample.py", line 97, in voxel_size, File "downsample.py", line 49, in down_sample sparse_pcd, cubics_ids = open3d.voxel_down_sample_and_trace( AttributeError: module 'open3d' has no attribute 'voxel_down_sample_and_trace'

If I use

sparse_pcd, cubics_ids = dense_pcd.voxel_down_sample_and_trace(voxel_size, min_bound, max_bound, False)

then

Processing: bildstein_station1_xyz_intensity_rgb Num points: 29697591 Num points after 0-skip: 9476296 Traceback (most recent call last): File "downsample.py", line 100, in voxel_size, File "downsample.py", line 53, in down_sample sparse_pcd, cubics_ids = dense_pcd.voxel_down_sample_and_trace(voxel_size, min_bound, max_bound, False) ValueError: too many values to unpack (expected 2)

But if I use:

sparse_pcd = cubics_idsdense_pcd.voxel_down_sample_and_trace(voxel_size, min_bound, max_bound, False)

it works, but I'm not sure were cubics_ids are stored. I know that sparse_pcd[0] is the down sampled point cloud, but I can't interpret sparse_pcd[1] and sparse_pcd[2].

VMohinder avatar Jun 17 '21 19:06 VMohinder

It seems that cubics_ids is stored in sparse_pcd[1].

VMohinder avatar Jun 17 '21 19:06 VMohinder

sparse_pcd, cubics_ids, some_list = open3d.geometry.PointCloud.voxel_down_sample_and_trace(pcd, voxel_size, min_bound, max_bound, approximate_class=False) this works for me

rekharchandran avatar Aug 18 '22 11:08 rekharchandran