> ## Content Index > Fetch the complete content index at: https://blog.topo-grapher.com/llms.txt > Use this file to discover other available public pages before exploring further. # LiDAR Versus Photogrammetry Elevations for BIM - URL: https://blog.topo-grapher.com/lidar-versus-photogrammetry-elevations-for-bim/ - Published: 2026-10-04T07:48:05.000Z - Updated: 2026-10-04T07:48:05.000Z - Author: Aki Olafsson A terrain model can look convincing and still be wrong where the project needs it most: at kerbs, drainage swales, retaining edges and the ground beneath trees. The choice between **LiDAR versus photogrammetry elevations** determines whether your BIM site model represents bare earth, visible surfaces, or a mixture of both. For architects and landscape teams, this is not a remote-sensing debate. It affects how confidently you set finished floor levels, test accessible routes, understand existing gradients and coordinate with civil information. Both methods can produce useful XYZ data, but they do not see a site in the same way. ## LiDAR versus photogrammetry elevations: the core difference LiDAR measures distance directly. A sensor emits laser pulses and records the time taken for each pulse to return. Airborne LiDAR surveys create a three-dimensional [point cloud](https://topo-grapher.com/learn/elevation-point-cloud.html?ref=blog.topo-grapher.com), with each point carrying eastings, northings and elevation. Where the source data has been classified, points can be separated into ground, vegetation, buildings and other surface types. Photogrammetry derives elevation from photographs. Software identifies matching features across overlapping aerial images and reconstructs their three-dimensional position. The resulting point cloud or raster surface can be highly detailed, particularly where there is clear texture, good image overlap and stable lighting. The practical distinction is simple. LiDAR may record several returns from a single laser pulse, allowing some pulses to pass through gaps in vegetation and reach the ground. Photogrammetry can only reconstruct what the camera can see. A leafy canopy, dense scrub or deep shadow may therefore become part of the model rather than revealing the terrain below. This does not make LiDAR automatically better. It makes it better suited to a different question. ## What each method actually represents Before importing any elevation file into Revit, Rhino or SketchUp, establish whether it is a [digital terrain model](https://topo-grapher.com/learn/dtm-vs-dsm.html?ref=blog.topo-grapher.com) or a digital surface model. A digital terrain model, often called a DTM, aims to represent bare earth. It is normally produced from classified LiDAR ground points, with buildings, trees and other objects removed. This is usually the more useful starting point for site grading, cut-and-fill studies, drainage direction and landscape levels. A digital surface model, or DSM, represents the first visible surface. Roofs, tree canopies, walls and sometimes vehicles can all appear in the elevation. Photogrammetry commonly produces DSM-style data, although post-processing can remove objects where supporting information and clear imagery are available. Consider a wooded slope behind a proposed housing plot. A photogrammetric surface may show a smooth but elevated blanket of tree crowns. That can be useful for visual context, overshadowing studies or understanding canopy height. It is not a reliable terrain surface for laying out a path or calculating a drainage fall. Classified LiDAR ground points are usually the appropriate source for that task. Conversely, on an open paved urban site with sharp kerbs, visible retaining walls and recent aerial capture, photogrammetry may describe surface detail very effectively. If the design question concerns what is physically exposed at the time of capture, a surface model can be valuable. ## Accuracy is more than a headline figure Elevation accuracy is often reduced to a single number, but that number does not tell the full story. Vertical accuracy depends on the survey method, sensor calibration, flight height, image quality, ground control, processing workflow, terrain cover and the datum used. LiDAR is generally dependable for terrain analysis when it has sufficient point density, sound ground classification and documented vertical accuracy. Lower-density national LiDAR can be entirely suitable for early feasibility and site massing, while a dense project survey may be needed before setting construction levels or designing detailed drainage. Photogrammetry can achieve excellent results in open, well-textured areas, especially when captured with rigorous control. Yet its quality can fall away on uniform asphalt, water, reflective roofs, deep shadows and vegetation. A smooth-looking photogrammetric mesh is not evidence of precise ground elevation. Interpolation can conceal gaps and artefacts. For both methods, check four items before treating the data as design-grade: - The capture date, because new earthworks, resurfacing or demolition can make accurate data obsolete. - The horizontal and vertical coordinate reference system, including the vertical datum. - The stated point spacing or ground sampling distance, which affects the smallest feature the model can represent. - The source’s documented accuracy and, for LiDAR, its classification method. A model can be geometrically tidy yet vertically offset by metres if its datum is misunderstood. That is a coordination problem, not a modelling problem. Confirm units, coordinate system and elevation reference before moving points into the project environment. ## Vegetation changes the decision Vegetation is the clearest reason to prefer LiDAR for many site models. Leaf-off LiDAR acquisition can provide substantially better access to terrain beneath deciduous trees than imagery-based reconstruction. It will not create perfect ground data through every woodland canopy, but it offers a defensible route to a bare-earth model when classified ground returns are available. Photogrammetry is strongest when the ground is visible. On a clear field, quarry, beach, construction site or paved square, it can produce detailed surface information that is visually intuitive and useful for rapid assessment. In dense woodland or overgrown land, it may model the vegetation convincingly while offering little evidence of the soil surface beneath it. There are edge cases. Tall grass, thick bracken and low shrubs can affect LiDAR ground classification too. A ground point beneath vegetation is not automatically a perfect terrain point, particularly where pulse density is low. Treat national datasets as informed context unless their specification and project risk justify more confidence. ## Point density and BIM performance More points are not always better inside a BIM model. A high-density point cloud may contain detail far below the scale of the design decision, while making a Revit Toposolid slow to regenerate or a SketchUp terrain mesh difficult to edit. Choose spacing to match the model’s purpose. Early massing and option studies often work well with a simplified grid that retains major breaks in slope. For paths, terraces, accessible gradients and local drainage, retain more points around critical level changes and built edges. Do not apply one uniform density simply because the source file contains it. A practical workflow is to preserve the original source data, then create a project-ready XYZ export at a manageable interval. Keep breaklines, survey spot levels and civil design data separate where they control a specific edge or grade. Terrain points describe the existing surface; they should not be expected to resolve every kerb line or underground drainage constraint. Topo-grapher is designed around this distinction: define the site boundary, generate terrain from available authoritative elevation sources, then download editable coordinates for the modelling environment rather than working from a visual web terrain alone. ## Selecting the right elevation source by task For concept design on an open site, either method may be suitable if the surface is current, the datum is understood and the model is used at an appropriate scale. Photogrammetry can be especially useful where visual surface character matters alongside elevation. For woodland sites, rural masterplanning, flood-path thinking and preliminary grading, classified LiDAR is normally the stronger starting point because it is more likely to provide a bare-earth terrain model. Confirm whether the dataset is ground-classified rather than assuming that all LiDAR outputs are terrain data. For detailed design, neither public LiDAR nor general aerial photogrammetry should quietly replace a commissioned topographic survey. Existing services, boundary evidence, kerb geometry, threshold levels and construction set-out require project-specific verification. Public elevation data helps teams begin earlier and ask better questions; it does not remove the need for a survey where risk and procurement demand one. ## A reliable workflow for terrain modelling Start by defining the area needed for design decisions, not merely the building footprint. Include the approach road, likely drainage outfall direction, neighbouring thresholds and enough surrounding land to understand slope continuity. A site clipped too tightly can create misleading edges in the model. Next, identify the source type and intended surface. Use bare-earth LiDAR ground data for terrain and grading studies. Use photogrammetric surface data where roofs, canopy or visible hardscape are relevant, and label it clearly so the model is not mistaken for ground. Then verify coordinates before import. Ensure the XYZ file uses metres or the project’s agreed unit convention, retain real-world coordinates where software and team standards allow, and record the horizontal and vertical reference. In Revit, establish shared coordinates deliberately before placing [a Toposolid](https://topo-grapher.com/topography-for-revit.html?ref=blog.topo-grapher.com). In Rhino and Grasshopper, account for large coordinates when precision or display performance becomes an issue. Finally, inspect the model against known features. Look for improbable spikes, flat patches beneath trees, stair-stepped interpolation and sudden datum shifts. Check a few elevations against survey control or trusted mapping where available. A five-minute review before modelling can prevent hours of redesign around a false slope. The useful question is not which technology wins. Ask what surface you need to design against, what the site conceals, and how much certainty the next decision requires. Choose the elevation source that answers that question, then keep its limits visible in the model.