uDocumentViewer
Geospatial.laz

LAZ viewer

ready
About LAZ

Compressed LIDAR point cloud.

LAZ is LASZip-compressed LAS: the same point records, but the X/Y/Z and attribute streams are arithmetic-coded chunk by chunk (50,000 points per chunk by default). We parse the LAS public header out of the buffer, then hand the rest to laz-perf — a WASM port of Martin Isenburg's reference laszip implementation — and pull point records out one at a time. Decompression is done on the main thread but yields back to the event loop every fifty thousand points so the UI stays responsive and the toolbar can update a "chunk K of M" progress indicator. Output is fed into the shared point-cloud viewer, where you can toggle between elevation, intensity, classification, and RGB colouring (when the point format includes RGB) and orbit/pan/zoom freely. Files above ~2M points are uniformly subsampled for display — the toolbar shows both the raw and shown counts.

FAQ
How is LAZ different from LAS?
LAZ is LAS run through LASZip — a lossless arithmetic coder tuned for point clouds. Files are typically 5×–10× smaller than the equivalent LAS but carry exactly the same coordinates, intensity, classification, and RGB.
Will it open LAZ 1.4 with the new extended point formats?
Yes. laz-perf handles point data record formats 6–10 (LAS 1.4 extended) as well as the legacy 0–5 formats, and our header parser reads the LAS 1.4 64-bit point count.
How long does decompression take?
A million points typically decode in a second or two on a modern laptop; ten million is closer to fifteen seconds. The progress indicator in the toolbar reflects laz-perf's 50,000-point chunk count.
Why don't I see RGB?
Only point data formats 2, 3, 5, 7, and 8 carry RGB. If your file is format 0, 1, 4, 6, 9, or 10 there are no per-point colours — switch the toolbar to elevation, intensity, or classification.
Is the .wasm fetched from a CDN?
No. The laz-perf WebAssembly module is bundled with this site and loaded directly from your browser. The point cloud never leaves your device.
More Geospatial