LiDAR Point Cloud file formats like ply, pcd, npy, npz, hdf5, binary, las, laz, and txt are very common in popular point cloud datasets like Toronto3D, Trimble, Toyota PCD datasets, ScanNet, ShapeNet, Sun RGBD, A2D2Audi Autonomous Driving Dataset, KITTI, USGS 3DEP, ModelNet40, Semantic3D, ModelNetC, ShapeNetC, ScanObjectNN and many more. Such variety in data formats sometimes makes it inconvenient to handle point cloud datasets. Visualization is an important step in machine learning/computer vision. This many file formats don’t make point cloud visualization any easier. In this tutorial, I’ve talked about some easy ways to visualize the most common file formats we see in many popular point cloud datasets in computer vision using Open3D and PPTK python packages. I hope you find this video helpful.
The list of file formats covered here is below, with references to the popular datasets they are found in.
ply (Toronto3D)
pcd (Trimble, Toyota PCD datasets)
npz, npy (ScanNet, ShapeNet, Sun RGBD, A2D2Audi Autonomous Driving Dataset)
las, laz (USGS 3DEP)
hdf5 (ModelNetC, ShapeNetC, ScanObjectNN)
binary (KITTI)
txt (ModelNet40, Semantic3D)

✅
Subscribe to the Channel    / @lightscameravision  

Code & Data https://github.com/LightsCameraVision...

Chapters:
The chapters are created using two pieces of information (point cloud file format python package).

0:00 Intro & Motivation
2:43 PLY Open3D
5:13 PCD Open3D
5:52 NPY Open3D
6:39 Las|Laz PPTK
8:40 HDF5 PPTK
9:47 Binary PPTK
10:30 txt PPTK & Open3D
11:25 Conclusion