Importing Point Clouds in Blender: A Point Cloud I/O Guide

What will you learn in this guide?

Learn how to use Blender’s Point Cloud I/O add-on to import scan data, check a point cloud’s appearance and attributes, and export it in a suitable format. Instead of converting files to meshes, the add-on uses Blender’s native PointCloud objects. This lets you work with points, colors, and other scan data using Geometry Nodes or shaders.

For architecture and interior design projects, this workflow can bring point data from laser scans or other sources into a scene so you can review it alongside an existing model, create data-driven visualizations, or export it in another format. However, the sources do not guarantee tasks such as turning a point cloud into a regular building component, validating measurements, or ensuring seamless exchange with other software. Before transferring data, check the file format and how the target application handles its attributes.

Requirements and compatibility

The Blender Extensions page lists E57, PLY, LAS/LAZ, PCD, XYZ, and PTS as supported formats for Point Cloud I/O. The add-on description says point colors, normals, intensity, classification, LiDAR return numbers, and custom fields in the file can be imported as point attributes. The attributes available depend on the source file and format.

The sources do not specify the minimum Blender version required by the add-on. Before installing, check whether it appears under Get Extensions in Blender. Some dependencies bundled with the add-on are provided as Python 3.13 wheels for macOS arm64, Linux x86_64, and Windows x86_64. Intel Macs are not supported out of the box; the sources note that the relevant wheels may need to be built from source on these systems. The add-on page does not list a price.

1. Install the add-on and prepare your file

In Blender, go to Edit > Preferences > Get Extensions. Search for Point Cloud I/O and install the add-on from the list. The source describes this as the add-on’s installation method and says there is no need to install dependencies manually with pip or download them from the internet.

Before importing, identify the format of your file. For example, LAS/LAZ is used for LiDAR data, while E57 files may contain multiple scans. For XYZ files, columns are interpreted automatically based on their number. Understanding these differences helps prevent the data from being assigned to the wrong fields in the scene.

2. Import the point cloud

Choose the import command for your format from the File menu. The add-on provides separate import options for E57, PLY, LAS/LAZ, PCD, XYZ, and PTS. For example, use File > Import > PLY Point Cloud (.ply) for PLY, or File > Import > LAS/LAZ Point Cloud (.las, .laz) for LAS/LAZ.

Each scan in an E57 file can be imported as a separate PointCloud object; you can combine scans using the Merge Scans option. XYZ imports do not require a header: three columns are interpreted as position, four as position and intensity, six as position and RGB, and seven as position, intensity, and RGB. A nine-column layout can also include RGB and normal components. For PTS files, the initial count line is read, and the add-on determines the field layout based on the row width.

If the imported cloud is not visible, check the coordinates, particularly for LAS/LAZ data with large coordinate values. Center on Origin is enabled by default in the add-on; it subtracts the minimum coordinate value to position the cloud near the origin. The original offset is stored on the object as las_origin_offset and is added back during export. This option is designed to help keep large coordinate values within the precision range of float32.

3. Check the appearance and attributes

After import, the add-on can create a Principled BSDF material driven by the point cloud’s color or normal. It also sets an initial point radius based on the cloud’s bounding box and density. These settings can make points visible as soon as the file opens, but you should check the result against the scale of the scene and the density of the scan.

In the 3D Viewport, press N to open the sidebar, then switch to the Point Cloud tab. Here you can see the point count and available attributes, and adjust the on-screen point size with the radius slider. The panel also includes ÷10, ÷2, Auto, ×2, and ×10 quick settings. It is not limited to objects newly imported with the add-on; it also works with selected PointCloud objects created by other methods.

Check attribute names and types before using them in Geometry Nodes or shaders. For example, color, intensity, and classification data serve different purposes, and not every source file will contain all of them. According to the add-on description, LAS classification can be imported from uint8 data as an INT attribute. If an expected field is missing from the workflow, check for it before exporting as well.

4. Check the format and potential data loss when exporting

Export the cloud using the appropriate File > Export command. Version 0.6.0 of the add-on adds a Source Attributes section to export dialogs. For channels supported by the format, you can select the source attribute, leave the selection on Auto, disable it with None, or assign another compatible attribute. According to the release notes, name matching now ignores differences in capitalization and punctuation, and some common alternative names are also recognized.

The same version adds warnings before export for attributes that will not be carried over to the file. These warnings help identify unexpected data loss, but testing the file in the target application is still the safest check. The add-on’s own PLY exporter differs from Blender’s built-in Stanford PLY exporter. Blender’s manual describes the built-in PLY workflow in terms of mesh data and mesh-specific options; for point clouds, make sure you use the add-on’s PLY Point Cloud option.

Important limitations: E57 exports do not write normals. Although an imported normal attribute can be preserved within Blender, it is not written back to an E57 file. LAS exports use a fixed scale of 1 mm, and positions are quantized to millimeter precision. LAZ compression can be enabled with the Compress (LAZ) option in the export dialog. Consider the project scale, target format, and attributes that need to be preserved before exporting.

Common mistakes and next steps

• Choosing the wrong PLY tool: Blender’s built-in Stanford PLY tool creates a mesh, while the add-on’s Point Cloud PLY tool creates a PointCloud object.

• Not seeing the points: Adjust the radius in the Point Cloud tab of the sidebar. The add-on sets a default radius automatically, but results can vary with scale.

• Missing attributes: Check the Source Attributes selections and data-loss warnings in the export dialog.

• Assuming E57 normals will be written back: The add-on does not export this field to E57; consider another transfer method if necessary.

• Overlooking precision with large-coordinate LAS/LAZ files: Account for the behavior of Center on Origin and the fixed millimeter scale on export.

Make your first attempt on a copy of the original project file. Then import a small sample cloud and test its appearance, attributes, and export to the target format. For architecture and interior design teams, the add-on offers a way to bring scan data into a Blender scene and use it in a visualization workflow; however, measurements, coordinate systems, compatibility, and attribute transfer should be verified for each project.

Sources and licensing

Adapted based on the Point Cloud I/O add-on page on Blender Extensions and the Stanford PLY document in the Blender manual. The Blender manual is licensed under CC BY-SA 4.0; its content has been adapted into Turkish. The add-on page does not specify a separate license for this content.