Point
Renders a standalone LAS/LAZ point cloud on the globe. Loading and parsing happen off the render thread; the first frame captures the Globe and georeferences the cloud into GPU-resident chunks, then every frame enqueues them through the engine's existing point path (DrawableTilePoints + Ogc3dTilesPointsProgram + PointCloudContent) — the same size-attenuated round splats the 3D-Tiles .pnts reader draws.
Georeferencing: geographic and WGS84/UTM coordinate systems are detected from the file's GeoKey VLR and converted automatically. For any other CRS, set projection to map native (x, y, z) to a globe Position; otherwise the coordinates are assumed geographic and a warning is logged.
Compressed .laz files need a LazChunkDecoder installed via earth.worldwind.formats.las.LasDecoderRegistry — loadLas fails clearly if it's missing.
Properties
How the cloud's Z is interpreted. AltitudeMode.ABOVE_SEA_LEVEL (default) treats it as orthometric height — how nearly all LiDAR ships — and adds the geoid offset so it shares the globe's datum and rests on terrain (terrain is elevation + geoid.getOffset; a raw orthometric Z would float by the geoid separation). AltitudeMode.ABSOLUTE treats Z as already-ellipsoidal. Ground-relative modes don't apply — a cloud's coordinates are absolute.
Metres to raise (or lower) the whole cloud, on top of altitudeMode. Applied live at draw time along the local up axis, so a change takes effect on the next redraw (the caller triggers it) with no reload. Nudge up to clear coarse terrain, correct a residual datum mismatch, or lift clear of the surface.
Optional native-CRS → globe Position mapping. When set it OVERRIDES the file's detected CRS, so it can reproject a coordinate system the engine doesn't know or re-place/re-base the cloud (e.g. drop it onto a flat globe). Applied per point.
Cascade-shadow participation. Standalone clouds default to none.