Drone LiDAR surveys that measure the ground beneath the trees

On hard, open surfaces, our Zenmuse L2 surveys typically achieve ±30–50 mm vertical RMSE when processed against a local GNSS base and checked on independent ground checkpoints. Under vegetation, accuracy depends on how many pulses reach the ground: where coverage is good the results approach open-ground figures, and under dense evergreen cover they degrade, which the report states. The manufacturer’s figure of 4 cm vertical at 150 m is a laboratory result. We report what the checkpoints on your site actually show.

It sees through the gaps. Each laser pulse can return several echoes, so some of the energy reflects from leaves and branches while some reaches the ground. Deciduous woodland in winter gives excellent ground coverage. Dense conifer plantations, ivy and thick bramble let far fewer pulses through. We assess the vegetation at scoping and tell you plainly what ground density to expect.

LiDAR measures distance directly with a laser, so it works in low light and records the ground beneath vegetation along with thin objects such as wires. Photogrammetry reconstructs geometry from overlapping photographs. On open, hard surfaces it gives sharper imagery and slightly better accuracy at a lower cost. Many of our projects combine the two, with LiDAR for the terrain and photogrammetry for the visual record.

The main cost drivers are area or corridor length, the required point density, vegetation complexity, the number of deliverables and whether CAD feature extraction is needed. Vegetation complexity matters because it drives classification effort. On heavily vegetated sites, classification and editing, not flying, is often the largest share of the cost. We provide a fixed quote against an agreed specification.

For terrain modelling and contours on most engineering projects, 100–200 points per square metre is ample. Vegetation analysis, overhead line clearance and detailed structure extraction benefit from 300 points per square metre or more. Specifying far more than you need adds processing time without improving the ground model, so we recommend a density matched to the purpose.

Point clouds are delivered as LAS 1.4 or compressed LAZ by default, with E57 available for BIM and reality capture software. Surfaces come as GeoTIFF rasters, LandXML TINs or XYZ grids, and linework as DWG, DXF or SHP. Everything is on OSGB36 British National Grid and ODN unless you specify otherwise.

Yes, and corridor work is one of the strongest uses of drone LiDAR. Flying near overhead lines or operational railways requires the asset owner’s agreement and additional risk controls, which we plan and document as part of the job. LiDAR returns from water are weak or absent, so along rivers and coasts we survey the banks and structures rather than the water surface or bed.

LiDAR is an active sensor, so it keeps working on dull days and in light too low for photogrammetry. We still don’t fly in rain, fog or heavy mist, which scatter the laser and shorten its range, and we restrict flying in strong wind to protect data quality. Colourising the point cloud with the onboard camera does need reasonable daylight.

For typical sites up to around 50 hectares, the classified point cloud and terrain model are usually delivered within five to seven working days of capture. Heavily vegetated sites, long corridors and CAD extraction add time, and we confirm the delivery date in the quote.