Drone-Based Photogrammetry and Aerial LiDAR for Topographic Mapping

Both methods turn data captured from the air into 3D terrain models. Photogrammetry reconstructs the surface from overlapping photographs, using software to match features across images. Aerial LiDAR fires laser pulses and measures their return time, directly recording millions of 3D points, including some that reach the ground through vegetation. Photogrammetry is cheaper and gives color imagery; LiDAR costs more but works better under tree cover and in low-texture areas. Many projects use one, the other, or both.

Part 1: Why Drones for Topographic Mapping

• Faster than ground survey: hundreds of hectares per day versus a few

• Dense data: millions of points instead of scattered spot levels

• Safer access to steep slopes, quarries, floodplains, and active sites

• Repeatable surveys for progress tracking and volume checks

• Lower cost than manned aircraft for small to medium areas

• Outputs feed directly into CAD, GIS, and BIM

Ground survey is still needed for control, checking, and fine detail such as kerbs, inverts, and building corners.

Part 2: How Photogrammetry Works

1. Flight planning: the drone flies a grid at set height with high overlap.

2. Image capture: a calibrated camera records geotagged photos.

3. Feature matching: software finds the same points across many images (Structure from Motion).

4. Camera positions and sparse cloud: the software solves where each photo was taken and builds a rough 3D point set.

5. Dense matching (Multi-View Stereo): produces a dense point cloud.

6. Products: mesh, Digital Surface Model, orthomosaic, and contours.

Key limitation: photogrammetry sees the surface. Where trees or crops hide the ground, it measures the canopy, not the terrain.

Part 3: How Aerial LiDAR Works

LiDAR (Light Detection and Ranging) uses a laser scanner, a GNSS receiver, and an inertial measurement unit (IMU).

1. The scanner emits pulses, often hundreds of thousands per second.

2. Each pulse reflects off surfaces and returns; the time gives the range.

3. GNSS and IMU record the sensor’s exact position and orientation.

4. Software combines range, position, and angle into georeferenced 3D points.

5. Multiple returns: one pulse can produce several returns, from canopy top, branches, and finally ground.

6. Ground points are classified and used to build a Digital Terrain Model (DTM).

Part 4: Comparison

Typical accuracy (well-controlled): photogrammetry commonly reaches about 2 to 5 cm horizontal and 3 to 8 cm vertical; drone LiDAR commonly about 2 to 5 cm vertical on open ground. Results depend on altitude, sensor, control, and processing, and should be verified by check points.

Part 5: Equipment

Platforms

• Multirotor: precise, hover capable, shorter endurance

• Fixed-wing and VTOL: longer endurance, large areas

Photogrammetry payload

• Mechanical-shutter camera with a good lens (global or mechanical shutter reduces distortion)

• Optional RTK/PPK for accurate image positions

LiDAR payload

• Scanner, GNSS, IMU, and often a camera

• Mounted on a heavier multirotor or VTOL

Positioning

• RTK (real-time kinematic): corrections during flight

• PPK (post-processed kinematic): corrections applied after flight; robust where radio links are weak

• Ground control points (GCPs) and checkpoints surveyed with GNSS or total station

Part 6: Planning the Mission

Part 7: Ground Control and Accuracy

• Place GCPs evenly across the site, including edges and different heights, commonly 5 to 10 for a modest site and more for large or corridor projects

• Use separate checkpoints not used in processing to report true accuracy

• Use stable, well-marked targets visible in imagery, or natural features for LiDAR checks

• Tie the survey to the project coordinate system and vertical datum, and confirm the geoid model

• RTK/PPK reduces the number of GCPs needed, but does not remove the need for checks

Report accuracy by RMSE of checkpoints, horizontally and vertically.

Part 8: Data Processing Workflow

Photogrammetry

1. Import images and positions

2. Align photos

3. Add GCPs and optimize

4. Build dense cloud

5. Classify ground (where possible)

6. Generate DSM, DTM, orthomosaic, and contours

7. Quality check against checkpoints

LiDAR

1. Process GNSS/IMU trajectory (PPK)

2. Generate and georeference the point cloud

3. Calibrate and align flight strips (boresight adjustment)

4. Clean noise

5. Classify ground, vegetation, buildings, and other classes

6. Generate DTM, DSM, canopy height model, and contours

7. Validate against checkpoints

Common software: Pix4D, Agisoft Metashape, DJI Terra, RealityCapture (photogrammetry); DJI Terra, TerraSolid, LiDAR360, and similar tools (LiDAR); Civil 3D, QGIS, and ArcGIS for deliverables.

Part 9: Products and Deliverables

Part 10: Applications in Civil Engineering

• Road and rail alignment: corridor mapping for route selection and design

• Earthwork volumes: cut and fill, stockpiles in quarries and mines

• Site grading and drainage design: accurate DTMs for hydrology and flood studies

• Dams, canals, and reservoirs: capacity and slope monitoring

• Landslide and slope monitoring: repeat surveys detecting movement

• Construction progress: compare as-built against design

• Utilities and power lines: LiDAR captures wires and clearances

• Land development and cadastral work: boundary and area mapping

• Mining and quarry: production tracking and safety

• Forestry and environment: canopy height and biomass (LiDAR)

Part 11: Choosing the Right Method

Combining both: LiDAR for the bare ground, photogrammetry for color and visual context, merged into one dataset.

Part 12: Limitations and Risks

• Photogrammetry fails under dense canopy, in shadows, and on reflective or featureless surfaces

• Wind, rain, and poor light degrade results

• LiDAR cannot penetrate very dense vegetation, so ground points may be sparse; ground classification needs review

• Datum and geoid errors cause systematic height offsets

• Poor control produces warped models with good-looking but wrong geometry

• Large datasets need strong computers and storage

• Regulations, no-fly zones, and permissions can limit operations

• Water surfaces give unreliable results; LiDAR near-infrared is absorbed by water

• Accuracy claims need independent checkpoints

Part 13: Quality Control Checklist

• Coordinate system and datum confirmed

• Control points surveyed and well distributed

• Independent checkpoints reserved

• Flight parameters match accuracy specification

• Weather and light suitable

• Image or point quality reviewed after flight

• Processing report examined

• RMSE values meet the specification

• Ground classification manually checked

• Contours and surfaces compared with ground-surveyed spot levels

• Deliverable formats and metadata supplied

Part 14: Common Mistakes

• Skipping ground control or checkpoints

• Using too little overlap or flying too fast, causing gaps and blur

• Flying in changing light or strong wind

• Ignoring the vertical datum or geoid model

• Treating a DSM as a DTM in vegetated areas

• Using automatic ground classification without review

• Not planning for battery and storage needs

• Overstating accuracy without independent verification

• Flying without proper permissions

Conclusion

Drone photogrammetry and LiDAR have changed topographic mapping from sparse spot levels to dense, repeatable 3D models. Photogrammetry is an economical, flexible way to map open sites and produce imagery, while LiDAR earns its higher cost where vegetation, wires, or featureless ground defeat photographs. In both cases, the quality of the result depends less on the drone than on good planning, solid ground control, correct datum handling, and independent checks. Accuracy figures here are typical rather than guaranteed, so define the required tolerance up front, verify it with checkpoints, and follow local aviation and survey regulations.