It is one of the questions I am asked most often, usually in the first five minutes of a call. My honest answer is that the site decides, not the brochure. Sometimes that means talking a client out of LiDAR they do not need. Sometimes it means explaining why the cheaper option would give them the tops of the trees instead of the ground.
Most drone survey enquiries eventually arrive at the same question: should this be flown with LiDAR or with a camera? Both methods produce a dense, georeferenced 3D model of the ground. Both can achieve survey-grade accuracy when they are properly controlled. But they collect data in fundamentally different ways, and that difference decides which one is right for your site.
How each method works
Photogrammetry
Photogrammetry reconstructs 3D geometry from overlapping photographs. The aircraft flies a grid with typically 70–80% forward overlap and 60–70% side overlap, and processing software matches thousands of features between images to calculate the position of each point on the surface. The output is a dense point cloud, a surface model and an orthomosaic: a geometrically corrected aerial image in which every pixel sits in its true map position.
Because it depends on matching visible features, photogrammetry only measures what the camera can see. Where the ground is covered, it measures the top of the cover. Where surfaces have little texture, such as still water, fresh snow, uniform tarmac or glossy roofs, matching becomes unreliable.
LiDAR
LiDAR (light detection and ranging) measures distance directly by timing laser pulses reflected from the surface. Combined with a precise position from GNSS and orientation from an inertial measurement unit, each return becomes a 3D point. A sensor such as the DJI Zenmuse L2 records up to five returns per pulse, so a single pulse can register the tree canopy, intermediate branches and the ground beneath.
LiDAR does not depend on light or texture, so it works on low-contrast surfaces and in flat light. It does not, however, produce the same quality of imagery. The L2 carries an integrated 20 MP RGB camera used to colourise the point cloud, but when a client needs a sharp orthomosaic for design or reporting, we usually fly a dedicated photogrammetry mission as well.
Vegetation: the deciding factor on most sites
Vegetation is the single most important factor when choosing between the two. LiDAR does not see through solid leaves; it passes through the gaps between them. Under a woodland canopy, a proportion of pulses reach the ground, and the more open the canopy, the more ground points you get. In winter, with deciduous trees out of leaf, ground returns increase considerably.
- Mature deciduous woodland in summer: LiDAR typically yields 5–20% ground returns, enough for a reliable ground model at 1 m resolution in most cases. Photogrammetry measures the canopy.
- Hedgerows and scrub: LiDAR usually captures ground levels on both sides and often beneath. Photogrammetry gives the top of the hedge.
- Long grass and crops: both methods measure some way above true ground. LiDAR performs better but can still sit 50–150 mm high in dense crop; timing the survey after harvest or cutting helps both methods.
- Dense evergreen conifers or bramble thickets: even LiDAR struggles. Expect gaps and plan for supplementary ground survey.
Accuracy compared
On open, hard ground with a properly designed control network, both methods can deliver vertical accuracy in the region of 15–30 mm RMSE, verified against independent checkpoints. The differences show up in specific conditions rather than in headline figures.
| Surface | Photogrammetry (vertical RMSE) | LiDAR (vertical RMSE) |
|---|---|---|
| Hardstanding, tarmac, concrete | 15–25 mm | 20–30 mm |
| Short grass, bare earth | 20–35 mm | 20–35 mm |
| Under open tree canopy | Not measurable (canopy only) | 40–100 mm |
| Steep faces and embankments | 20–40 mm with oblique imagery | 25–45 mm |
| Low-texture surfaces (water, snow, uniform roofs) | Unreliable | Reliable except on water |
Photogrammetry often edges ahead on hard surfaces because its accuracy scales with ground sample distance (GSD), and a 45 MP full-frame camera flown at 60–90 m achieves a GSD of 8–12 mm. It also defines edges well: kerb lines, painted markings and building corners are easy to see and digitise in a sharp orthomosaic. LiDAR's per-point noise is typically higher, but its ability to reach the ground under cover means it is often the only method that can deliver a valid ground model at all.
Whatever the method, the accuracy you can rely on is the accuracy measured against independent checkpoints, not the sensor's datasheet figure. Our article on drone survey accuracy explains how to read and specify those numbers.
Deliverables: what each method produces best
| Deliverable | Photogrammetry | LiDAR |
|---|---|---|
| Orthomosaic | Excellent, 10–30 mm GSD | Basic, from integrated camera |
| Bare-earth DTM on open ground | Good | Good |
| DTM under vegetation | Poor | Good |
| Digital surface model (DSM) | Good | Good |
| Textured 3D mesh | Excellent | Limited |
| Overhead lines and thin structures | Poor | Good |
| Hard detail for CAD (kerbs, markings) | Excellent | Adequate, better with imagery |
| Stockpile volumes | Excellent | Good |
Two items stand out. Overhead power lines and telecom cables are too thin for photogrammetry to reconstruct, but LiDAR records them as distinct returns, which makes it the natural choice for vegetation encroachment and clearance checks on energy corridors. Conversely, textured 3D meshes for visualisation, heritage recording or façade work are the preserve of photogrammetry.
Cost
LiDAR carries a higher cost per project, for three reasons. The sensor and associated GNSS equipment are more expensive to own and maintain. Processing requires trajectory computation, strip alignment and point classification, which is more specialist work. And separating ground from vegetation to produce a reliable DTM takes manual quality control that scales with how vegetated the site is.
As a rough guide, a LiDAR survey of a given site is often 30–60% more expensive than a photogrammetry survey of the same area and accuracy. On an open site, that premium buys very little. On a vegetated site, it buys the only valid ground model available, and is usually far cheaper than the alternative of a walked survey through woodland.
“The most expensive survey is the one you have to fly twice. I would rather tell a client they do not need LiDAR than sell them a premium that buys nothing.”
Other site conditions
- Light and weather: photogrammetry needs consistent daylight and is affected by hard shadows; LiDAR can fly in flat light and early or late in the day. Neither method flies in rain, and both are limited by wind.
- Water: neither method measures the bed of a river or pond through the water column with standard drone sensors. LiDAR returns from water surfaces are sparse and noisy; photogrammetry fails to match on moving water. Plan wading or boat survey for bed levels.
- Steep and vertical faces: photogrammetry with oblique imagery handles cliffs, quarry faces and façades well. LiDAR captures them too, but nadir flights leave shadows behind overhangs.
- Urban sites: photogrammetry is usually preferred for its imagery; LiDAR helps where trees line the streets.
When to combine both
On a large proportion of our projects, the answer is both. A typical greenfield housing site has open fields, where photogrammetry gives an excellent surface and orthomosaic, crossed by hedgerows and ditches, where LiDAR is needed for ground levels. Flying both missions from the same aircraft platform on the same day, tied to the same control network, adds a modest cost to the job and removes the weaknesses of each method.
The merged dataset gives the design team a LiDAR-derived ground model where it matters, photogrammetric detail for hard features, and a high-resolution orthomosaic as a base plan for every consultant on the project.
A quick decision guide
- Is more than about 10% of the area you need levels for under trees, hedges or dense scrub? Choose LiDAR, or LiDAR plus photogrammetry.
- Do you need a high-resolution orthomosaic, textured mesh or crisp hard detail? Include photogrammetry.
- Are you surveying overhead lines, pylons or vegetation clearance? Choose LiDAR.
- Is the site open, hard-surfaced or a stockpile yard? Photogrammetry is usually the most cost-effective option.
- Are there low-texture surfaces such as large uniform roofs or snow cover? Favour LiDAR, or reschedule.
- Not sure? Send us a site boundary and a description of what you need to design or decide. We will recommend a method and explain why.
What to include in your brief
You do not need to specify the sensor. Specify the outcome: the area, the deliverables, the required accuracy and how it will be verified, the coordinate system and the date you need the data. A competent survey contractor should then propose the method and justify it. Be wary of any proposal that recommends photogrammetry for a wooded site without explaining how ground levels under the canopy will be obtained.
Related services
