# Point Cloud

TunnelCAD supports importing and processing 3D point cloud data for automated cross-section extraction. This eliminates the need for manual profile digitization by sampling measured tunnel geometry directly from laser scan data.

## Workflow overview

The typical point cloud workflow in TunnelCAD follows these steps:

```{image} ../_static/images/pointcloud_workflow.svg
:alt: Point cloud workflow diagram
:width: 100%
```

1. **Import** raw point cloud data (LAS, LAZ, E57, PLY, PTS, or XYZ format) into the project.
2. **View** the point cloud in the **Point Cloud** canvas tab, which renders the data using OpenGL.
3. **Open the Section Viewer** to interactively inspect point cloud slices at alignment stations.
4. **Adjust filter parameters** (angular binning, radius limits, elevation cutoffs) to isolate the tunnel surface.
5. **Extract** the filtered points into measured cross-section line data at one or more stations.
6. The extracted sections join the tunnel's cross-section list and can be used for volume calculations, clearance analysis, and reporting.

## Supported formats

| Format | Extension | Description |
|--------|-----------|-------------|
| LAS | `.las` | ASPRS LIDAR data exchange standard |
| LAZ | `.laz` | Compressed LAS (requires liblaszip library) |
| E57 | `.e57` | ASTM E2807 standard for 3D imaging data |
| PLY | `.ply` | Polygon File Format -- simple vertex data |
| PTS | `.pts` | Leica point cloud format (text: X Y Z Intensity R G B) |
| XYZ | `.xyz` | Plain text format (X Y Z per line, optionally with RGB) |

## IQPC cached format

When you import a raw point cloud, TunnelCAD converts it to the **IQPC** (IQsoft Point Cloud) format -- an optimized binary representation with a spatial octree index. The IQPC file is cached alongside the project and reloaded on subsequent sessions without re-parsing the original file. This makes repeated work on the same dataset substantially faster.

IQPC files can also be shared between projects via **Point Cloud > Attach (.iqpc)...**.

## Point Cloud canvas tab

The **Point Cloud** tab in the main canvas area renders all loaded point clouds in 3D using OpenGL. You can orbit, pan, and zoom to inspect the data before opening the Section Viewer for extraction.

```{note}
On Linux, point cloud import requires two shared libraries in the application directory:

- `libE57CWrapper.so` -- for E57 format support
- `liblaszip.so` -- for LAZ compressed format support

These are not needed on Windows, where the equivalent DLLs are bundled with the application.
```

```{toctree}
:maxdepth: 1

importing
section-viewer
attaching
```