A point cloud is a set of spatial points. A mesh is a connected surface representation. Both can be useful in a projection workflow, but they are different assets with different review requirements.
A point cloud records spatial observations
Laser scanning produces points representing surfaces visible from each capture position. Points can include attributes such as color or intensity depending on the equipment and workflow. Areas that cannot be observed are not automatically measured.
Individual scan positions need to be aligned into a common reference. This is registration. A registered point cloud should be reviewed for coverage and the quality of the relationships between scans before downstream work depends on it.
Point density describes how many measurements represent an area. It does not, by itself, establish positional accuracy. Capture method, registration, reference control, surface conditions and validation all matter.
A mesh describes connected surfaces
A mesh is built from vertices, edges and faces. When derived from a point cloud, it turns sampled measurements into a surface representation. This can make geometry practical for a 3D content or visualization workflow.
The conversion can involve interpolation, smoothing, hole filling or simplification. Those operations have consequences. A smooth surface may lose a sharp scenic edge; an automatically filled gap may represent an assumption rather than a measured feature.
Keep the measured reference available. Document which surfaces are modeled, which are omitted and where interpretation has been introduced.
A production asset needs a purpose
The most detailed mesh is not always the most usable. A content pipeline may need a manageable polygon count, consistent normals and a specific UV strategy. A fabrication comparison may need critical interfaces preserved. An AV planning model may prioritize boundaries, context and reference features.
These requirements should be agreed before processing. A request for “all the data” can produce a large handoff without clarifying what the receiving team actually needs.
Questions to ask before delivery
- What information was directly measured, and what was derived?
- Which areas were not captured or modeled?
- What are the units, origin, axes and spatial reference?
- What simplification or interpolation was applied?
- Which formats and versions does the receiving application support?
- Who owns UV mapping, content templates and final calibration?
- How will the representative file be checked before release?
The handoff to projection
Measured geometry can support planning and a content model. It does not alone establish projector calibration. Projection platforms can also use camera-based reconstruction and mesh deformation, so the best approach depends on the complete workflow.
Discuss the spatial data service or use the capture planning guide to define a brief.
Technical references
- Disguise OmniCal overview: point clouds, mesh deformation and calibration.
- Leica entertainment workflows: spatial capture in entertainment production.