A one-pilot operation based in Slough, Berkshire, flying the same saved mission over the same site month after month for construction progress monitoring — a dated visual record for site meetings and programme evidence, processed in-house into a pack you can compare period to period. Not survey-grade. Scope, formats and the limits of the kit are all stated below; every job is quoted per site.
Scroll the table sideways →
| Service | You receive | Stated accuracy |
|---|---|---|
| Construction progress pack | Orthomosaic JPG, 15–30 labelled stills, month-to-month pair, PDF pack | Visual record, not survey-grade |
| Roof & building inspection | Labelled 4K stills + PDF defect sheet, numbered and located on a plan | Visual condition record, not a structural survey |
| Site overview & marketing stills | Graded stills, optional short cut | — |
| Topographic survey / stockpile volumesHired kit | GeoTIFF, point cloud, DXF linework, written accuracy statement | Quoted per job with the accuracy spec attached |
| Thermal / solar inspectionHired kit | Radiometric imagery, per-panel fault map | Quoted per job |
| Cinematic — half day | Aerial and FPV to an agreed shot list, raw or graded | — |
| Cinematic — full day | As above, full day on location | — |
| Edit package | Graded horizontal master plus vertical cutdowns | — |
Every job is quoted per site, in writing, with a unit and a scope attached — prices exclude VAT. Travel within one hour of Slough is included. A site address and approximate size are enough to get a figure rather than a question.
Who the monthly pack is built for: main contractors and project managers, and QS / commercial teams. Also worked with, quoted per job: groundworks contractors (earthworks progress and a dated record that travels between sites), survey practices (subcontract capture on a saved mission — you keep the client, I fly the grid), quarry operators (per-visit visual record of stock and cell movement only, not audited volumes, which need hired RTK and a different accuracy statement) and FM / estates teams (roof and facade condition records, usually one-off rather than a monthly series).
On bare, hard ground, repeat surveys of the same site agree with each other to within 0.24 m for 90% of the surface and 0.31 m for 95%.
This is repeatability, not accuracy: it measures how well the method agrees with itself between visits, not agreement with a coordinate reference frame, which needs independent checkpoints withheld from processing. Measured on a validation site on 4 August 2026 by flying it twice and comparing. Bare and hard ground only — vegetation and tree canopy are several times worse. Percentiles, not RMS: the error distribution has heavy tails. Change is quoted relative to the baseline survey, never as absolute levels or heights AOD.
Standard visits fly a DJI Mini 3 Pro, which has a rolling shutter, and the validation flight above was nadir-only, so dome deformation across the site is unquantified in that dataset. Georeferencing is GNSS-only — no RTK, no PPK, no ground control. These outputs are not survey-grade, must not be used for setting out, boundary determination, or measurement for payment, and I will not put a number on a stockpile from them. Where a job needs that, I hire an RTK aircraft and set ground control.


These are my own captures, flown on real sites to the same specification as a client brief — same aircraft, same saved mission, same pack at the end. Everything above is genuine output, not a mock-up.