Guides08 Sept 2026 2 min read

UAV Photogrammetry for Engineering Survey: From Flight Planning to Orthomosaic

The full controlled photogrammetric workflow — GSD, overlap, RTK/PPK, ground control, bundle adjustment and accuracy assessment — explained for engineering use.

By GLONAS Engineering Consultant
Share
UAV Photogrammetry for Engineering Survey: From Flight Planning to Orthomosaic

The workflow

Workflow
  1. 01Flight planning
  2. 02Capture
  3. 03RTK/PPK + GCPs
  4. 04Bundle adjustment
  5. 05Dense reconstruction
  6. 06DSM/DTM
  7. 07Orthomosaic
  8. 08Accuracy assessment

Ground sample distance (GSD)

GSD is the real-world size of one pixel on the ground. It sets the resolution limit of every downstream product. A 2 cm GSD means no object smaller than ~2 cm is reliably resolved. GSD is determined by flight altitude, camera focal length and sensor pixel size; choose it to match the smallest feature the survey must capture.

Overlap and sidelap

  • Overlap (along-track): typically 80% — enough that each feature appears in many images for robust matching.
  • Sidelap (cross-track): typically 60–70% — enough to tie flight lines together.

Too little and matching fails or gaps appear; too much and processing cost rises without quality gain.

Flight altitude

Altitude sets GSD and influences coverage rate. Lower altitude gives finer GSD but more images and longer processing; higher altitude covers more ground but coarsens resolution. The choice is a trade against the required accuracy and the site size.

RTK/PPK and ground control

  • RTK/PPK records the camera position to centimetre-level at capture, reducing the number of ground control points needed and strengthening the block geometry.
  • Ground control points (GCPs) are independently surveyed points visible in the images that anchor the block to a known coordinate system and datum. They are what make the output metric.

A common engineering setup uses RTK/PPK for camera positions plus a small number of check points to validate accuracy. GCPs remain the defensible anchor for survey-grade work.

Camera calibration and image quality

A calibrated camera (known focal length, principal point, distortion) improves bundle adjustment. Mechanical shutters avoid rolling-shutter distortion; good exposure and sharp focus keep matching reliable. Image quality is not optional — it propagates into every product.

Bundle adjustment

Bundle adjustment simultaneously solves camera positions, orientations and 3D point coordinates to minimise reprojection error across all images. It is the geometric heart of the pipeline; its residuals indicate block quality.

DSM, DTM and orthomosaic

  • DSM — digital surface model, including objects on the surface (buildings, vegetation).
  • DTM — digital terrain model, the bare-earth surface.
  • Orthomosaic — the geometrically corrected, mosaicked image, differentially rectified against the DSM so it is measurable like a map.

Accuracy assessment

Accuracy is not claimed; it is measured. Independent check points (not used as control) are compared against their position in the orthomosaic/DSM. The residuals — RMSE in X, Y and Z — are the documented accuracy of the deliverable. Without this step, "accuracy" is an assertion.

The deliverable

A survey-grade orthomosaic, a DSM/DTM and a documented accuracy report, tied to a stated datum and coordinate system. That is the product an engineering decision can rely on — and the standard a controlled UAV photogrammetry campaign is built to meet.

Tags
UAV PhotogrammetryEngineering SurveyGSDRTKOrthomosaic