Point Cloud Engineering Concepts

A point cloud represents sampled 3D positions from photogrammetry, LiDAR, or another reconstruction process. Each point can carry coordinates and optional attributes such as RGB color, intensity, or classification.

LAS is a standard point cloud exchange format. LAZ is a compressed form that preserves point data more efficiently. ZIP packages are sometimes used to group point cloud files and supporting metadata.

Potree organizes large point clouds into an octree so a browser can stream the level of detail needed for the current camera view. This is why large datasets can be inspected in a web viewer without downloading every point at once.

3D measurements are generally more direct than image measurements because points already exist in scene coordinates. However, they still depend on reconstruction quality, coordinate scale, point density, alignment, and whether the clicked points represent the physical surface the reviewer intends to measure.

Verticality compares a selected axis against gravity or the scene's vertical axis. The output can include height, drift from plumb, lean angle, and lean direction. It is useful for towers, poles, mounts, and linear members.

Correction geometry is derived from reviewer-selected points. It can make sparse structures easier to interpret, but it should always be labeled as an inferred or fitted object rather than raw point evidence.