The fastest way to choose a reality-capture method is to ask what someone must do with the result. A textured mesh for a public presentation, a controlled point cloud for topographic measurements, a BIM reference for visible piping, and a weekly construction record are four different deliverables. The sensor vocabulary matters after the output, tolerance, access, and environment are clear.
LiDAR measures distance with emitted laser light. “3D laser scanning” is a broad field term that includes LiDAR instruments and workflows, from static terrestrial scanners to mobile systems using simultaneous localization and mapping, or SLAM. Photogrammetry reconstructs geometry from overlapping images. Each can produce a point cloud. That shared output does not make the capture behavior or evidence equivalent.
A deliverable-first comparison
| Decision | LiDAR / laser scanning tends to help | Photogrammetry tends to help |
|---|---|---|
| Surface appearance | Intensity and optional color may be available | Image texture is a core strength |
| Low light | Active ranging can capture geometry without scene illumination | Images need suitable exposure and visible texture |
| Plain or repetitive surfaces | Range observations can still describe geometry | Reconstruction may struggle without image features |
| Dense vegetation | Multiple returns may provide observations through canopy gaps | Images mainly describe visible surface |
| Highly visual mesh | Often needs a color workflow or separate imagery | Well-planned imagery can support detailed texture |
| Confined mobile route | SLAM can support capture where GNSS is unavailable | Needs sufficient overlap, lighting, and visual features |
These are tendencies, not acceptance criteria. A particular camera, LiDAR, lens, trajectory, altitude, surface, and processing chain can change the outcome. The project team should require a test dataset whenever a critical feature sits near the limit of the proposed method.
If the deliverable is a measured point cloud
Begin with the required coordinate system, accuracy class, density or spacing, coverage, classification, and file format. The USGS 3D Elevation Program illustrates how specific a LiDAR deliverable can become: quality levels combine positional accuracy with nominal pulse spacing or density, and the current collection specification adds requirements for project extent, returns, intensity, voids, and distribution. A building or plant project will use different numbers, but it benefits from the same habit of declaring measurable properties.
Mobile SLAM LiDAR makes field collection exceptionally fluid. The operator can walk through connected spaces while the system estimates a trajectory from observations of the surroundings. That freedom carries a planning duty. Drift, weak geometry, moving objects, and incomplete loop connections can affect the result. Control and independent checks remain essential where the output will support survey or dimensional decisions.
Static laser scanning remains attractive where fixed viewpoints, carefully registered setups, and high-detail observations of a defined scene match the task. Mobile and static capture can also work together: a mobile pass supplies broad context, while static stations or another measurement method strengthen critical zones.
If the deliverable is a textured model
Photogrammetry earns a strong look when surface color and recognizable texture lead the brief. A planned image network needs substantial overlap, stable exposure, sharp images, varied viewpoints, and scale or control. Reflective, transparent, uniformly colored, moving, or deeply shadowed surfaces require special care. The crew should inspect image sharpness and coverage in the field, because a blurred set cannot recover detail through patient processing.
For an exterior facade, roof, aggregate pile, or accessible site, imagery may produce both geometry and a presentation-ready surface. A drone can extend viewpoints, subject to FAA rules, airspace, site permission, weather, visibility, and crew safety. Under Part 107, commercial small-UAS operations require a certificated remote pilot or direct supervision by one, along with applicable operating and airspace requirements.
If the deliverable is a BIM reference
The model author rarely needs every captured point at maximum density. That person needs a cloud that opens reliably, sits in the agreed coordinates, contains required surfaces, and carries enough information to identify modeled elements. Autodesk documents ReCap as a bridge for registering, cleaning, organizing, and converting point clouds into RCP or RCS data used by Revit and other products. The handoff still needs shared coordinates, units, segmentation, and modeling scope.
No capture method automatically creates an accurate BIM. Modeling introduces interpretation: which pipe diameter to use, how to represent insulation, whether a wall is modeled as-built or idealized, and what level of detail is justified by the visible data. The contract should separate measured cloud accuracy from model tolerance and authoring decisions.
If the deliverable is change evidence
Repeatability outranks maximum novelty. The team should preserve route maps, control, coordinate system, naming, clipping extents, and comparison settings. LiDAR is compelling for fast geometric coverage in stable coordinates; photography is excellent for readable visual context. Many change programs should capture both, provided files can be related by location and time.
A comparison also needs a baseline policy. Temporary materials, people, vehicles, vegetation, wet surfaces, and moved equipment can generate apparent changes. The analyst needs thresholds, exclusion rules, and a way to trace each reported item back to source data. A red-versus-blue deviation image may look decisive while mixing registration error with actual movement.
Questions that settle the choice
- What will the user measure, model, inspect, or communicate?
- Which tolerance applies to the final product, and how will it be tested?
- Does the site have light, texture, GNSS, safe access, and stable surfaces?
- Is color central, helpful, or irrelevant?
- Which areas are hidden from the proposed path or viewpoints?
- What coordinate frame and formats must downstream software receive?
- Can a pilot area be captured and accepted before the full mobilization?
The answer may be a hybrid. LiDAR can supply controlled geometry, imagery can supply texture and visual evidence, and conventional survey observations can supply control and checks. A good proposal explains what each source contributes, how they will be aligned, and where uncertainty remains. Method labels then become useful shorthand instead of substitutes for a deliverable.