LAS Point Cloud Filter
LocalFilter LAS point clouds by classification/intensity and export PLY/CSV/XYZ/OBJ/JSON — full parse, no subsampling, locally in your browser. Pro feature.
Drag files here, or click to browse
Supports LAS
Choose fileWhat is it? LAS Point Cloud Filter selects points by classification or intensity range and exports the filtered set to PLY/CSV/XYZ/OBJ/JSON. The file is parsed in full (no pre-filter subsampling), so every matched point is kept; the result line reports matched/total counts — e.g. extracting the ground class (ASPRS 2) for a DTM deliverable. Pro feature (point-cloud export/filter share one Pro key).
TL;DR: Filter LAS by class/intensity and export the matched set — full parse, no pre-filter subsampling, runs locally in your browser. Pro feature.
How to use
- Open a .las file — it parses in full and the loaded line shows the total count.
- Set the classification or intensity filter (advanced options: bounding box, export cap).
- Click Filter & export — the result line reports how many points matched out of the total.
FAQ
Which classifications can I filter?
The fixed set shown on the page: Never Classified, Unclassified, Ground, Low/Medium/High Vegetation, Building, Noise, Water and Road. Points with other class codes cannot be selected here (leave all boxes unchecked to keep them).
Is it local?
Yes.
Is it free?
No — LAS Point Cloud Filter is a Pro feature (point-cloud export/filter share one Pro key); on the Free plan this page shows an upgrade prompt.
What about compressed or very large clouds?
For compressed LAZ, decompress first; full parsing is memory-bound, so for huge clouds use the cloud-based Filter Point Cloud tool (account required, no pre-sampling there either).
How do I extract only the ground points?
If your file already carries ASPRS classifications, tick Ground (class 2) — every matched point is kept and the result line shows matched/total counts. This tool selects points by their stored classification; it does not run a ground-classification algorithm, so raw or never-classified scans need classifying first (CloudCompare, PDAL or similar).
How is this different from the class filters in the LAS Viewer?
The viewer's class legend and intensity range only change what is drawn on screen. This tool writes the matched subset to a real file (PLY/CSV/XYZ/OBJ/JSON) you can hand downstream — display filtering discards nothing.
Examples
- Drop in a .las → pick classifications → export
Supported formats
Limitations
- Uncompressed LAS only.
- Files are parsed in full — no 4M-point subsampling before filtering, and all matched points are kept; tens of millions of points are memory-bound (use the cloud filter for huge files).
- The optional export cap evenly samples the matched set (the filtered result), never the raw file; the result line shows matched/total counts.
- Filters are simple ranges (classification + intensity), plus a bounding box and export cap under advanced options.
- Empty result raises an error — relax the filters.
- CSV/PLY/JSON keep intensity and classification (plus RGB when present; JSON carries the attributes and is larger than the old coordinates-only output); XYZ/OBJ are coordinates-only.
- Runs locally in your browser; nothing is uploaded.
Privacy
Privacy first: this tool runs entirely in your browser — your input and files never leave your device, and everything is cleared when you close the page.