KITTI & Velodyne viewer
Inspect LiDAR scans and labels in 3D in your browser—no install. Use it to sanity-check data, teach how KITTI-style folders fit together, or step through a short driving sequence with optional camera images next to the point cloud.
What you can do here
- One frame — Load a single
.bin(x, y, z, intensity). Addcalibandlabel_2to draw 3D boxes in the LiDAR frame. - Whole sequence — Load KITTI training folderpicks a directory like KITTI's
trainingtree. You get sorted frames, a scrubber, and play/pause so you can see how the scene changes over time and optionally see matchingimage_2photos. - Try instantly — Load bundled demo loads sample points and labels with no files of your own.
Coloring: Distance highlights structure by range; Intensity shows return strength. Built with React and Three.js.
Single scan: use the file inputs on the right—good for one-off .bin files.
Sequence: Load KITTI training folder and pick your training folder (or a parent whose paths still contain velodyne/, calib/, label_2/, optional image_2/). You can scrub or play through frames to see motion and, with camera files, compare 2D and LiDAR for the same id.
Keys when a sequence is loaded: ← → step, space play/pause (not in text fields). Recenter camera refits the 3D view; scrubbing keeps your angle so you can compare neighbors.
No point cloud loaded
Use Load bundled demo, Load KITTI training folder, or a single .bin.