LiDAR vs. photogrammetry: two ways to build a point cloud, and when each wins

A point cloud is a site rendered as millions of measured points, each with an X, Y, Z — the raw material for surface models, volumes, and 3D design checks. Two technologies produce them from a drone, and the sales pitches for each tend to obscure a fairly simple set of trade-offs.

How photogrammetry does it

Photogrammetry takes the same overlapping photos used for an orthomosaic and, by matching features across images, triangulates a 3D position for an enormous number of pixels. Because every point comes from a photo, every point carries true colour, and the clouds are very dense. The catch is that it can only reconstruct what the camera can see and distinguish: blank, featureless surfaces (fresh concrete, still water, glass) give it little to match on, it needs decent light, and the results take processing time.

How LiDAR does it

LiDAR fires laser pulses and times their return, measuring distance directly — no feature matching, no dependence on texture or lighting, and results that are essentially ready when the flight ends. Its signature advantage on real sites is vegetation: pulses slip through gaps in the canopy and some reach the ground, so LiDAR can model terrain under trees that a camera simply can't see. The costs are hardware (survey-grade LiDAR sensors are a significant investment), sparser point spacing, and no native colour unless a camera is paired with it.

What a side-by-side actually shows

Pix4D published a controlled comparison scanning the same structure with photogrammetry and with the LiDAR built into a phone. Their photogrammetry cloud was markedly denser and reproduced edges more crisply, because a laser only returns a point where it happens to hit, while a camera reconstructs every visible pixel. On a 3.18 m object they measured about 4 mm of error from photogrammetry versus 12–47 mm from the phone LiDAR depending on where the points were taken, and the phone sensor's useful range was about 5 m. That test is handheld, consumer-class LiDAR — survey-grade drone LiDAR performs far better — but the underlying trade-offs it illustrates hold at every price point: density and edge fidelity versus direct, texture-independent measurement.

Choosing for a construction or facility project

• Open earthworks, stockpiles, pads, paving, roofs and façades → photogrammetry. Dense, colour, sharp edges, and the same flight produces your ortho.
• Wooded corridors, pipeline and power-line rights-of-way, bare-earth terrain under canopy → LiDAR.
• Textureless or reflective surfaces (large fresh slabs, water, glass) → LiDAR, or photogrammetry with ground control placed to anchor those areas.
• Large sites with both conditions → many teams fly both and merge; Pix4D's own tools, for example, fuse the two.