Photogrammetry and LiDAR can both create detailed 3D records, but they measure the world differently. The right method depends on whether the project prioritises measured geometry, rich visual surface information or both.
The fundamental difference is how the geometry is obtained.
LiDAR measures range by emitting laser energy and recording its return from visible surfaces. A terrestrial laser scanner collects large numbers of spatial measurements from each instrument position. Multiple positions can then be registered into a coherent point cloud.
Photogrammetry derives three-dimensional geometry by identifying corresponding features across overlapping photographs taken from different viewpoints. With sufficient image overlap and suitable geometry, processing can reconstruct camera positions, a point cloud, mesh and textured surface.
Both can produce point clouds and three-dimensional outputs, but a point cloud created by photogrammetry is not produced by the same measurement process as a laser-scanned point cloud.
There is no useful universal accuracy figure for either technology.
A professional terrestrial laser scanner can provide a strong basis for high-quality measured building work, but finished survey accuracy still depends on range, surfaces, network geometry, registration, control and QA. See Terrestrial Laser Scanning Accuracy.
Photogrammetric accuracy depends on factors including image resolution, camera geometry, overlap, focus, surface texture, calibration, scale and control.
For survey work, the important question is not whether one technology can produce an impressive-looking model. It is whether the chosen methodology provides sufficient verified geometry for the project's intended use.
Where dependable metric geometry is the priority, Geo Scan generally treats controlled survey measurements as the authoritative spatial reference rather than assuming image-derived geometry has equivalent performance.
Photogrammetry can capture extremely rich visual information because photographs are fundamental to the process.
It is particularly useful for:
The resulting textured model can communicate material, colour, deterioration and surface character in a way that geometric linework alone cannot.
Its success depends on the scene. Featureless, reflective, transparent, moving or poorly illuminated surfaces can be difficult, and capture must provide suitable image overlap and viewpoint geometry.
3D laser scanning is well suited to measured environments where geometry and spatial relationships are the principal requirement.
Typical strengths include:
LiDAR also has limitations. The scanner only measures surfaces visible from its positions. Occlusion can require many viewpoints, and difficult surface characteristics can affect observations.
The survey still needs deliberate network design, registration and checks. More points do not automatically mean more accuracy.
A laser-scanned point cloud can be colourised where suitable imagery is captured, giving each measured point RGB information. That can make the dataset much easier to interpret.
Photogrammetry goes further by using imagery to reconstruct and texture continuous surface representations. For heritage, visualisation and surface recording, this can be a major advantage.
This is one reason the technologies should not always be framed as competitors. Laser scanning can provide the measured spatial framework while photogrammetry contributes dense image-led surface information.
Geo Scan used this principle on The Calderstones, where laser scanning provided measured spatial reference and scale while photogrammetry recorded the stones as photorealistic three-dimensional surfaces.
A photogrammetric reconstruction can have internally consistent shape without automatically being positioned at the correct real-world scale or within the required survey coordinate framework.
Surveyed control can address this.
Clearly identifiable targets or control points can be coordinated independently and introduced into the photogrammetric processing. This can establish scale, constrain the model and provide points against which results can be assessed.
The appropriate arrangement depends on the subject, capture geometry and accuracy requirement. Control should be distributed to constrain the area of interest rather than concentrated where it is convenient to place.
Independent check information is also valuable because control used to build a solution is not the same thing as an independent verification of it.
Neither photogrammetry nor LiDAR can directly record a solid surface hidden behind another object.
For LiDAR, additional scanner positions are used to see around walls, equipment, furniture and building services. In a congested plant room or MEP survey, viewpoint diversity can be critical.
Photogrammetry likewise requires the surface to appear clearly in suitable overlapping images. A surface seen in only one photograph cannot be reconstructed in the same way as a feature observed from multiple useful viewpoints.
Good reality capture is therefore partly an exercise in visibility planning: understanding which surfaces matter and designing enough complementary viewpoints to record them.
The technologies can overlap, but they often lead naturally to different outputs.
| Photogrammetry | LiDAR / laser scanning | |
|---|---|---|
| Primary observation | Overlapping imagery | Direct range measurements |
| Common 3D outputs | Image-derived point cloud, textured mesh | Registered point cloud |
| Visual surface information | Excellent when imagery is suitable | Colourised points where imagery is captured |
| Survey geometry | Can be strong with appropriate capture and control | Strong basis for measured geometry with appropriate control and registration |
| Typical 2D outputs | Orthophotos / orthographic imagery | Plans, elevations and sections derived from point cloud |
| Common uses | Heritage, façades, visual surface records, aerial mapping | Measured buildings, engineering, MEP, CAD/BIM |
| Key scene dependency | Texture, lighting and image geometry | Line of sight, range, surface response and scan geometry |
The table describes typical strengths, not hard boundaries. Both methods can support a much wider range of applications when appropriately designed.
A hybrid approach is useful when the project needs both dependable measured geometry and richer visual surface information.
One workflow is to establish the spatial framework using terrestrial laser scanning and survey control, then capture high-overlap photography for selected surfaces or objects. The image-derived data can be scaled or coordinated to the measured reference.
This can be particularly useful for heritage recording, complex façades and other projects where a conventional point cloud or drawing does not communicate all of the visual information present on site.
The methods do not need to compete. The strongest workflow can use each technology for the information it captures best.
"LiDAR" covers more than static terrestrial laser scanners.
Mobile LiDAR and SLAM scanning can capture spaces rapidly while the sensor moves through the environment. This can be valuable for supplementary coverage and particular site conditions, but its error behaviour and control strategy differ from static terrestrial scanning.
Geo Scan's controlled mobile workflow treats mobile capture as an additional tool rather than assuming it replaces terrestrial laser scanning for every measured survey. Where tighter geometry is required, mobile data can be related to stronger static TLS or surveyed control.
That distinction is important when comparing photogrammetry with "LiDAR": the specific LiDAR platform and methodology matter.
Gaussian splatting is another image-led method for producing visually rich representations of real environments.
Its strengths are different from a controlled measured point cloud. A Gaussian-splat scene can provide compelling visual context and intuitive navigation without automatically becoming authoritative survey geometry.
Geo Scan is exploring controlled hybrid workflows in which visually rich image-led capture complements established measured data rather than replacing it where verified geometry matters.
This follows the same principle as photogrammetry and LiDAR: choose each dataset according to the job it is being asked to perform.
Start with the output and decision the survey needs to support.
Choose terrestrial laser scanning where dependable measured building geometry, complex interiors, engineering relationships or CAD/BIM production are central requirements.
Choose photogrammetry where high-quality visual surface information, textured models, orthographic imagery or image-led heritage recording are central requirements.
Consider both where geometry and visual surface information are important.
Then define the accuracy, coverage, control, deliverables and QA appropriate to the project.
Sometimes the most valuable dataset is not the one captured with the highest possible precision. It is the one that captures the right information, quickly, completely and at an appropriate level of accuracy for the task.