Tiny Online Tools logoTiny Online ToolssearchSearch tools…grid_viewAll Tools
Homechevron_right3D Toolschevron_rightPoint Cloud ProcessorPoint Cloud Processor

Point Cloud Processor

Crop, merge, downsample, denoise, transform and recolour LiDAR point clouds from LAS, LAZ, PLY, XYZ or PCD files.

upload_file

Click to browse or drag & drop files here

Select one or more point clouds — LAS, LAZ, PLY, XYZ or PCD

Accepted: .las,.laz,.ply,.xyz,.pcd

Pick several files to merge them into a single cloud. LAS and LAZ tiles are re-aligned against the first file so they do not stack on top of each other.
crop CropNoneBoxAxis rangeKeep the full extent of the cloud.grain DownsamplingNoneVoxel gridEvery Nth pointKeep every point that survived the crop.filter_alt Noise removalRemove statistical outliersMeasures the mean distance from each point to its k nearest neighbours and drops the points that sit too far out. Capped at 2,000,000 points — downsample first for anything larger.tune TranslateRotate (degrees)Applied in XYZ order, then baked into the coordinates — PLY and XYZ have no node transforms to carry.A non-uniform scale on survey data is almost always a mistake, so only a uniform factor is offered.ColourWrites the RGB values the source file carried, if it had any.Binary PLY is far smaller and loads faster; ASCII PLY and XYZ are human readable.Include colour in the exportXYZ carries RGB as three extra columns; some readers expect exactly three columns and will choke on six.infoSteps always run in this order: merge, crop, downsample, remove outliers, transform, recolour. Cropping first keeps every later step cheap, and the outlier test only makes sense on a cloud of roughly even density.

Similar Tools

Point Cloud to Mesh

Point Cloud to Mesh

Reconstruct a watertight surface from a LAS, LAZ, PLY, XYZ or PCD point cloud and export it as STL or GLB.

Point Cloud Viewer

Point Cloud Viewer

Open and explore LAS, LAZ, PLY, XYZ and PCD point clouds directly in your browser, free and without uploading anything.

Point Cloud Converter

Point Cloud Converter

Convert LAS to PLY, LAZ to XYZ, PCD to PLY and other point cloud formats free in your browser, with optional downsampling on the way through.

Mesh to Point Cloud

Mesh to Point Cloud

Sample the surface of a 3D model into a point cloud in your browser and export PLY or XYZ.

RAW Image Viewer

RAW Image Viewer

Preview camera RAW files (CR2, NEF, ARW, DNG, RAF, ORF, RW2 and more) by extracting the embedded JPEG preview.

Markdown to HTML

Markdown to HTML

Convert Markdown to HTML instantly with live preview.

Posterize Image

Posterize Image

Reduce the number of color levels in any image to create a bold posterized effect. Adjustable levels, runs entirely in your browser.

apps

More Tools

Browse our full collection of free online tools.

Six edits every scan needs, in the right order

A raw scan is rarely usable as delivered. It is too dense, it includes the tripod and the operator, it has noise floating above the roofline, it sits in a projected coordinate system nowhere near the origin, and it arrives split across tiles. This free online point cloud processor does all of that in one pass, in the browser, with nothing uploaded.

What it does

  • Merge several LAS, LAZ, PLY, XYZ or PCD clouds into one file
  • Crop to a box, or to a range on a single axis — expressed as a percentage of the cloud's own extent, so you can set it before you know the coordinates
  • Downsample with a voxel grid (uniform density) or every-Nth thinning (fast preview)
  • Remove statistical outliers using the mean distance to each point's k nearest neighbours
  • Transform — translate, rotate, scale, or centre on the origin or the ground plane
  • Recolour by elevation, by LiDAR intensity, or with one flat colour

Why the order is fixed

The pipeline always runs crop, then downsample, then outlier removal, then transform, then recolour. That is not arbitrary.

Cropping first makes everything after it cheaper, and the crop box is defined against the original bounding box, so it has to be evaluated before anything moves the data. Downsampling before outlier removal matters even more: the statistical test compares each point's neighbour distance against a global average, and that is only meaningful when density is roughly even — which is exactly what voxel downsampling produces. Recolouring runs last because an elevation ramp has to map the final height range.

Removing outliers without removing detail

For every point the tool measures the mean distance to its k nearest neighbours, then drops points beyond mean + ratio × standard deviation. Noise — dust, birds, mixed pixels at depth edges — sits further from its neighbours than surface points do, so it fails the test first.

The ratio is the knob that matters. Around 1.0 is aggressive and will eat thin structures like wires and foliage. 2.0 is a sensible default. 3.0 removes only gross blunders. The result panel reports how many points each stage removed, so you can tune it instead of guessing.

Merging tiles safely

LAS and LAZ files store coordinates in the hundreds of thousands of metres, and each one is recentred on its own bounding box when opened so rendering stays precise. Merging tiles naively would stack them on top of each other; this tool re-offsets every cloud against the first one instead, and reports the offset so you can restore the original coordinate system afterwards.