Surveyors have been arguing about this one for years, and the argument keeps getting more complicated. LiDAR scanners that once cost six figures are approaching the price of a good total station. Photogrammetry software has gotten smart enough that a competent operator can produce survey-grade outputs from overlapping drone imagery in an afternoon. But cheaper and faster doesn’t mean interchangeable. The physics of how each method captures the world are fundamentally different, and those differences dictate everything: what environments you can work in, what deliverables you can produce, and whether your data is actually telling you what you think it is.
What You’re Actually Capturing
Where Photogrammetry Earns Its Place
Where LiDAR Scanners Earn Their Cost
The Accuracy Numbers in Context
Choosing Based on What the Job Requires
The Investment Question for Canadian Firms
The Data Has to Hold Up. Choose Accordingly.
Most comparisons treat LiDAR and photogrammetry as two roads to the same destination. They aren’t, and conflating them leads to expensive mistakes.
Photogrammetry is a passive method. Survey drones equipped with high-resolution cameras record reflected ambient light across overlapping frames. Software identifies common points between those frames and triangulates their position in three-dimensional space through a process called structure from motion. The geometry is inferred, not measured. It exists because the algorithm found enough visual agreement across images to reconstruct it, which means it can only exist where the algorithm has something to work with.
LiDAR is an active method. The sensor emits its own energy: laser pulses fired at rates of hundreds of thousands per second. Each pulse travels to a surface, reflects, and returns. The system records the precise time of that round trip and converts it to distance. Nothing about that process depends on colour, texture, or available light. The sensor is measuring geometry directly, with its own illumination, against physics that don’t change between a clear July morning and a grey November afternoon in northern Ontario.
That distinction, inferred geometry versus measured geometry, is the root of every practical difference between the two methods.
On cooperative sites, photogrammetry produces exceptional results at a fraction of LiDAR’s cost. A subdivision under active construction, a gravel pit requiring stockpile volumes, a highway corridor with minimal canopy, these are environments where a well-executed photogrammetry mission with proper RTK or PPK correction delivers horizontal accuracies in the 1 to 3 centimetre range. The outputs are visually rich: textured meshes, coloured point clouds, and orthomosaics that engineers, project managers, and clients can read without specialist training.
That visual interpretability is worth more than people typically account for. A textured 3D mesh of a road cut or building pad communicates progress and conditions in a way a bare geometric point cloud does not. For documentation, client reporting, and stakeholder communication, photogrammetry gives you something immediately useful to everyone on the project team.
The limitations are real and worth stating plainly. Photogrammetry’s accuracy depends entirely on what the camera can see and how well the algorithm can match it across frames. Flat, homogenous surfaces, snow, standing water, bare ice, uniform concrete, provide insufficient texture for reliable point matching, and accuracy collapses. Shadows from steep terrain or low sun angles create noise, occlusions, and data voids. Vegetation is a hard ceiling: the camera resolves the canopy surface with precision and tells you nothing about what’s beneath it.
LiDAR scanners become the only viable option the moment a site stops cooperating with a camera, and in Canada, that threshold is crossed regularly.
Dense boreal forest, brush along pipeline corridors, creek valleys choked with willows, mixed-cover terrain across the Shield, photogrammetry in these environments produces a precise model of the top of the vegetation. It maps what the camera can see, which is not the ground. LiDAR pulses diffuse through gaps in canopy that no image-based sensor would detect. With sufficient pulse density, ground returns come back through cover that is effectively opaque to photogrammetry, making it possible to generate a bare-earth digital terrain model where a camera-based workflow simply cannot.
Lighting independence compounds this advantage in Canadian conditions specifically. Survey drones flying photogrammetry missions require adequate, consistent daylight, a resource that diminishes sharply through fall and winter at northern latitudes. LiDAR operates in complete darkness. It functions in the flat, diffuse light of an overcast December day that would compromise photogrammetric image matching. For projects with tight timelines that don’t accommodate waiting for ideal conditions, that operational flexibility has direct value.
Vertical accuracy is the third area where LiDAR holds a structural advantage. Photogrammetry’s vertical precision is inherently weaker than its horizontal, typically half to a third as strong, because nadir-facing cameras capture far less angular information about elevation than about planimetric position. That gap widens on terrain with significant relief, where varying distances to the ground produce uneven ground sampling distance and shadows obscure surface texture precisely where accurate elevation data is most critical. LiDAR measures elevation as directly as it measures anything else. The vertical accuracy floor is higher and far more consistent across terrain types.
Both methods can achieve centimetre-level accuracy. That statement is technically true and practically incomplete without the conditions attached.
Photogrammetry reaching 1 to 2 centimetre horizontal accuracy requires good ground control or reliable RTK/PPK correction, consistent and even lighting, overlap in the 70 to 80 percent range, and a site surface with enough texture variation to give the matching algorithm traction. All of those conditions have to hold simultaneously. In Canadian field conditions across a full calendar year, they frequently don’t all hold at once.
LiDAR reaching 2 to 3 centimetre vertical accuracy requires a calibrated sensor, quality IMU data, and proper georeferencing. It does not require cooperative weather, cooperative lighting, or a cooperatively textured surface. The accuracy floor is less dependent on conditions that are outside the operator’s control, which matters enormously when you’re committing to deliverable specifications on a project with a fixed deadline.
The decision framework isn’t about which technology is superior. It’s about which one matches the specific demands of the project in front of you.
Photogrammetry is the right call when:
LiDAR is the right call when:
It’s also worth understanding that hybrid workflows are increasingly standard practice, not an edge case. LiDAR for bare-earth geometry and photogrammetry for surface texture and visual context produces a dataset that is simultaneously precise and interpretable. At Bench-Mark, we work with clients across pipeline, mining, and environmental sectors who have attempted photogrammetry in vegetated terrain and returned with data that can’t support the deliverables they committed to. The cost of remobilization and recapture consistently exceeds what the LiDAR system would have cost in the first place.
LiDAR entry costs have dropped, but the gap with photogrammetry remains significant. Before committing to ownership, the questions worth working through honestly are:
At Bench-Mark, we help clients work through that evaluation with specifics, including hardware compatibility with the field software they’re already running.
Engineers, regulators, and clients don’t care which sensor collected the data. They care whether it’s accurate, complete, and defensible against the conditions of the site it came from. Choosing between LiDAR and photogrammetry on cost alone, or on what worked on the last project, is how firms end up remobilizing. Match the method to the environment, the deliverable, and the conditions, and in Canadian survey work, that match matters more than almost anywhere else.
