Choose your operating system:
Windows
macOS
Linux
The
LiDAR Point Cloud plugin
supports a variety of coloration and shading techniques and makes modification and visualization of point clouds quick and easy, even at runtime.
Performance
Performance is primarily governed by the Global Point Budget, which sets the maximum number of points that can be displayed at one time. Using a shared budget as opposed to a per-component budget allows for efficient rendering of many assets at the same time, as the system will select the best points from all of the visible components A higher point budget means higher quality, which means more costly performance. The global point budget saves VRAM and helps frame rates but does not reduce overall RAM usage.
On-Demand Streaming
When opening a Point Cloud, only the necessary header information will be loaded, while the actual bulk data is streamed on-demand. This results in fast asset loading as well as minimizing total RAM consumption.
The editor does not release memory of any assets automatically - you will need to re-launch after saving to take advantage of streaming those assets. This is applicable to both newly imported and upgraded legacy assets
Loading large numbers of point cloud files in the engine still requires a significant amount of RAM when parsing point cloud data and processing it as Unreal assets. For example, this public LiDAR data from the city of Montreal loaded in Unreal has the following performance indicators:
|
Individual LAS files |
~684 tiles for a total of 253 GB on disk |
|
Unreal RAM usage |
~3.5 GB for a budget of 1M points |
|
Frame Rate |
120 FPS, with a global budget count of 1M points |
|
Total Amount of Points |
Avg 14.3M points per 1km*1km tile x 621 files = ~8.9 Billion points |
Lower densities (where lower budget is selected) can be somewhat mitigated by increasing Point Size (uniform multiplication value for all points) and Point Size Bias (affects how the scaling is performed per LOD).
During situations where the budget for assets is small and the number of overall components is high, point cloud assets can disappear and pop-in. To prevent this, the LOD system actively attempts to allocate at least the minimum quality LOD for assets that are far away from the camera.
Importing Point Clouds
You can import new assets using either of the following methods:
-
Drag-and-drop a point cloud saved in a supported file format (XYZ, TXT, PTS, LAS) to your Content Browser.
-
Select Import from the Content Browser and navigate to the desired file.
On import, meters are converted to UU (x100). This can be modified in the Project Settings.
Exporting Point Clouds
Point Cloud Assets can be exported as ASCII or LAS files using the existing Unreal Export tool .
On export, UU are converted back to meters (x0.01). This can be modified in the Project Settings.
Editing Point Clouds
Toggling between Center and Original Coordinates wil re-center the point cloud data around its center. This is useful if the Point Cloud data has been defined in coordinates that are far away from the model's origin.
Individual points can be deleted, merged, hidden, and extracted.