Easy Ways to Visualize 3D Point Clouds: Common Datasets & Formats in Python

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 RGB-D, A2D2-Audi Autonomous Driving Dataset, KITTI, USGS 3DEP, ModelNet40, Semantic3D, ModelNet-C, ShapeNet-C, 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 RGB-D, A2D2-Audi Autonomous Driving Dataset) - las, laz (USGS 3DEP) - hdf5 (ModelNet-C, ShapeNet-C, ScanObjectNN) - binary (KITTI) - txt (ModelNet40, Semantic3D) --------------------- ✅👍📸 Subscribe to the Channel 👉 https://www.youtube.com/channel/UCXeI6BuHPYDd-Po9CZz-bnw?sub_confirmation=1 --------------------- Code & Data 👉 https://github.com/LightsCameraVision/Point-Cloud-Visualization --------------------- 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