A drone LiDAR topographical survey is normally specified to RICS Band D or Band E of Measured Surveys of Land, Buildings and Utilities, 3rd edition (RICS, 2014): ±10 mm or ±25 mm in plan at one sigma on hard detail. The accuracy actually achieved is measured against independent ground check points and stated in the survey’s accuracy report; it is never assumed from the sensor specification.
This is the long-form version of the answer on our LiDAR and laser scanning service page and topographic survey service page. It explains what the bands mean, what governs the accuracy a UAV LiDAR survey can reach, and how to tell a verified figure from an advertised one.
What does “accuracy” mean on a survey specification?
In UK practice the accuracy of a measured survey is expressed as a band from the RICS specification’s survey detail accuracy band table. Each band states a plan tolerance and a height tolerance at one sigma (68 per cent confidence); the two-sigma (95 per cent) value is double the one-sigma value. The bands most relevant to UAV work are reproduced below from the RICS band table (RICS, 2014, pp. 17 to 18), plan and hard-detail height only.
| RICS band | Plan (1 sigma) | Height, hard detail (1 sigma) | Typical use in the RICS table |
|---|---|---|---|
| A | ±2 mm | ±2 mm | Monitoring, precise setting-out, fabrication |
| B | ±4 mm | ±4 mm | High-accuracy engineering and measured building survey |
| C | ±5 mm | ±5 mm | Engineering setting-out, heritage |
| D | ±10 mm | ±10 mm | Engineering and topographic survey, determined boundaries |
| E | ±25 mm | ±10 mm | Topographic survey, measured building survey, PAS 128 QL-A verification |
| F | ±50 mm | ±50 mm | Lower-accuracy topographic survey, utility tracing |
| G | ±100 mm | ±50 mm | Topographic survey, boundary mapping, PAS 128 QL-B1 detection |
| H | ±250 mm | ±125 mm | Lower-accuracy topographic survey, national urban mapping, tree surveys |
Two points follow. First, a band is a statement about the whole deliverable, not about the best point in it. Second, the RICS table distinguishes hard detail (kerbs, structures, manhole covers, hardstanding) from soft detail (ground under grass, crops or scrub). A survey can honestly hold Band D on hard detail while the bare-earth model under vegetation sits in a looser band. A good specification and a good report both say which is which.
Which RICS band is a drone LiDAR survey specified to?
For topographic work, UAV LiDAR is normally specified at Band D or Band E on hard detail, with the bare-earth DTM beneath vegetation accepting the looser bands inherent in ground extraction under cover. That is the same range as well-controlled UAV photogrammetry; the difference is that LiDAR holds it across vegetated, shadowed or texture-poor ground where photogrammetry cannot form a surface at all.
Where a project needs Band A to C (deformation monitoring, structural verification, rail or highway setting-out), the tool is terrestrial laser scanning or a total-station traverse, not an aircraft. Our point cloud capture methods comparison sets out that split.
What determines the accuracy of a UAV LiDAR survey?
A LiDAR point is computed from the sensor’s position, its orientation and a laser range, so every one of those inputs contributes error. The sensor’s brochure figure describes only the ranging component under ideal conditions. The table lists what governs the accuracy that arrives in the deliverable.
| Error source | What governs it | How it is controlled |
|---|---|---|
| Sensor position | Quality of the GNSS solution for the aircraft trajectory | Post-processed kinematic (PPK) against a base station tied to OS Net; PPK is preferred to real-time correction because it uses the full observation set |
| Sensor orientation | Inertial measurement unit drift and boresight calibration between IMU and scanner | Calibration flights, cross-strips and strip adjustment against overlapping passes |
| Laser range and footprint | Flying height, beam divergence, incidence angle on slopes, surface reflectance | Flying height and line spacing chosen for the specification; overlapping passes on steep aspects |
| Point density | Pulse rate, ground speed, flying height, number of passes | Planned to give enough ground returns for a reliable DTM after classification |
| Absolute tie to the grid | Number, distribution and quality of ground control | GNSS-observed GCPs on OSGB36 and ODN, distributed across the whole envelope; see ground control points for drone mapping |
| Ground classification | Vegetation density, terrain roughness, algorithm settings | Automated filtering followed by manual editing where automation fails: embankment toes, cut faces, hedgerow lines |
| Datum transformation | Correct application of OSTN15 and OSGM15 | Applied in processing and stated on the deliverable; see OSGB36 drone survey workflows |
None of these has a fixed value that can be quoted in advance. The achieved accuracy is the sum of them on the day, on that site, which is why it has to be measured.
How is the accuracy of a drone LiDAR survey verified?
The verification method is the one that separates a survey you can build from and a survey that merely looks right: independent check points.
- Ground control points (GCPs) are surveyed targets fed into the strip adjustment. Their residuals show only that the solution fitted its own control.
- Check points are surveyed to the same standard, on the same datum, but withheld from the solution. The difference between the survey surface and each check point is the true error.
- The report gives the root mean square error (RMSE) in plan and height across the check points, plus the maximum residual, because the worst point governs fitness for a tight-tolerance use.
A worked example from our own work: on a 250 hectare solar and battery-storage site in Lincolnshire, surveyed by fixed-wing UAV LiDAR and photogrammetry together, 24 GCPs were established at a maximum spacing of 350 metres and eight independent check points were withheld. The check points returned a vertical RMSE of ±3.1 cm and a horizontal RMSE of ±2.3 cm against a ±5 cm project specification. That figure, not a brochure figure, is what the case study reports.
Our QA/QC methodology guide covers the check-point workflow stage by stage.
Is the ground under vegetation as accurate as hard detail?
No, and a report that claims otherwise should be questioned. On hard detail the laser returns from the surface being measured. Under grass, crop or canopy the ground surface is reconstructed from the subset of returns that reached it, so the DTM accuracy there depends on how many ground returns there were and how well they were classified. Check points observed on open ground do not prove the DTM under a hedgerow; a rigorous programme places some check points on soft ground and reports them separately. Our guide can drones survey through vegetation? covers what governs that.
What should the accuracy report contain?
- Number and distribution of GCPs and independent check points, shown on a plan
- Plan and height RMSE and the maximum residual, hard and soft detail separated where relevant
- The RICS band the result corresponds to, with the sigma stated
- Horizontal datum and transformation (OSGB36 via OSTN15) and vertical datum and geoid (ODN via OSGM15), or the site grid and its relationship to the National Grid
- Trajectory processing, strip adjustment and classification summary
- Processing software and version, flight dates, and the equipment used
If the report does not distinguish check points from control, treat its accuracy figure as a measure of internal consistency rather than of accuracy.
How Angell Surveys does this
Angell Surveys is regulated by RICS and specifies every UAV LiDAR survey to a named RICS band before flying. We fly a fixed-wing VTOL platform, the Wingtra Ray, with its LiDAR payload for large-area and corridor work, and terrestrial laser scanning where Band A to C is required. Ground control is observed by dual-frequency GNSS tied to OS Net, with independent check points withheld from every strip adjustment and their residuals reported in the deliverable. On a water company’s twin reservoir programme covering about 70 square kilometres, and on a renewable energy developer’s 250 hectare consenting survey, the datasets were accepted by the Planning Inspectorate and statutory consultees on the strength of that documented verification.
Related services
Sources
- RICS, Measured Surveys of Land, Buildings and Utilities, 3rd edition, November 2014 (reissued as an RICS professional standard, December 2023): survey detail accuracy band table.
- RICS, Earth Observation and Aerial Surveys, 6th edition, March 2023.
- Ordnance Survey, A Guide to Coordinate Systems in Great Britain, v3.6, 2020: sections 6.3 (OSTN15) and 6.4 (OSGM15).
- Angell Surveys case study: 250 ha solar farm and BESS topographic survey, Lincolnshire.
- Angell Surveys case study: Lincs and Fens reservoirs aerial survey.