Drone photogrammetry and mapping: from overlapping photos to defensible data

Software finds the same visual features in multiple photographs, estimates where each camera was, triangulates points in three dimensions, and builds a surface. From that surface it can generate an orthomosaic, point cloud, digital surface model, contours, mesh, or volume. A pretty stitched image is not automatically a map, and camera GPS is not automatically proof of accuracy.

The client question

What decision will this dataset support, and how accurate, complete, current, and interoperable must it be for that decision?

A beginner can learn on an ordinary RGB drone and public sample imagery. Selling measured boundaries, certified topography, design surfaces, or survey-grade claims can fall within state surveying or engineering practice. The durable entry path is visual progress mapping, planning-grade orthomosaics, or capture subcontracting under a licensed professional.

The output defines the job. Name the file, its acceptance test, who reviews it, and what it cannot prove before the aircraft launches.

Output 01

Orthomosaic

Used for

Site context, annotation, inventory, and repeat-date comparison.

Check

Coordinate system, complete coverage, and stated limitations.

Handoff

GeoTIFF, cloud web map, JPEG preview

Technical note+

A geometrically corrected image assembled from many photographs. It is useful for site context, annotation, inventory, and change comparison when its coordinate system and limitations are documented.

Output 02

Point cloud and surface models

Used for

3D inspection, DSM or DTM creation, and measurable surface work.

Check

Classification quality, gaps, and independently checked geometry.

Handoff

LAS/LAZ, GeoTIFF DSM/DTM, XYZ

Technical note+

A 3D set of measured points plus a DSM or DTM. A DSM includes trees and structures; a DTM attempts to represent bare terrain and requires classification and QA.

Output 03

Contours and volumes

Used for

Grade review, stockpile volumes, drainage context, and CAD handoff.

Check

Surface quality, base definition, units, and checkpoint error.

Handoff

DXF, SHP, GeoPackage, PDF report, CSV

Technical note+

Derived measurements that depend on a trustworthy surface, base definition, coordinate units, and checkpoint evidence. Report assumptions and measured error with the number.

Output 04

Accuracy and metadata report

Used for

Document how the dataset was made and where its limits are.

Check

CRS, datum, GSD, control, checkpoints, residuals, and gaps.

Handoff

PDF plus machine-readable project metadata

Technical note+

The part that makes the imagery defensible: CRS, horizontal and vertical datum, GSD, control source, checkpoint count, residuals, gaps, capture conditions, and limitations.

1. Step 1 of 6

   Define the decision and acceptance criteria

   Write the area of interest, output formats, coordinate reference system, vertical datum, target ground-sampling distance, delivery date, and accuracy language before planning a route.

2. Step 2 of 6

   Plan overlap, height, motion, and control

   A common starting point is roughly 80% front and 80% side overlap, adjusted for terrain, texture, obliques, camera, and software. Decide whether camera positions alone are enough or whether the project needs RTK, PPK, ground control, and independent checkpoints.

3. Step 3 of 6

   Capture and inspect in the field

   Lock an appropriate exposure strategy, avoid blur and major lighting changes, watch battery and wind margins, and check coverage before leaving. A missed strip is cheaper to refly on site than after processing.

4. Step 4 of 6

   Align, reconstruct, and clean

   Review camera alignment and tie-point quality before generating dense products. Remove obvious artifacts, classify ground only when the workflow supports it, and retain the original imagery and processing report.

5. Step 5 of 6

   Validate with evidence that was not used to fit

   Ground control points help shape the model. Checkpoints are withheld from adjustment and test the result. Report the actual residuals or RMSE rather than calling RTK output survey grade by default.

6. Step 6 of 6

   Package for the next tool

   Open exported data in QGIS, ArcGIS, CAD, or the client's platform. Confirm units, northing/easting, vertical reference, scale, nodata areas, and file naming before delivery.

Buy from the accepted output backward. A more expensive sensor cannot rescue a vague scope, weak method, invalid conditions, or missing review authority.

Capability 01RGB mapping aircraft

A supported aircraft with a sharp camera and repeatable automated flight is enough to learn. A mechanical shutter and RTK improve demanding work but do not replace checkpoints.

Capability 02GNSS and targets

Survey-grade control generally requires appropriate GNSS equipment, durable targets, a known datum, and a person competent to establish and evaluate control.

Capability 03Workstation and storage

Photogrammetry uses substantial RAM, GPU, disk, and upload capacity. Preserve source imagery, processing reports, deliverables, and project metadata in a recoverable archive.

Capability 04LiDAR

LiDAR measures laser returns and can sample ground through canopy gaps. It is not a magic see-through-trees sensor; trajectory, boresight, strip alignment, classification, and checkpoints still matter.

Flight planning, processing, GIS, asset analysis, client delivery, and job sourcing are different jobs. Confirm aircraft, controller, payload, operating-system, export, storage, and offline compatibility before paying for a subscription.

• DroneDeploy

  Role

  Capture, cloud processing, repeat-site comparison, and client sharing

  Use it when

  A team wants a managed web workflow and recurring site history.

• PIX4Dmatic / PIX4Dmapper

  Role

  Professional photogrammetry processing

  Use it when

  You need detailed processing controls, reports, and established geospatial exports.

• Agisoft Metashape

  Role

  Local, scriptable photogrammetry

  Use it when

  You want workstation processing and control over alignment, dense reconstruction, and export.

• WebODM / OpenDroneMap

  Role

  Open photogrammetry workflow

  Use it when

  You are learning, self-hosting, or need an inspectable lower-cost stack.

• QGIS

  Role

  GIS inspection, measurement, styling, QA, and packaging

  Use it when

  Always. It is a free way to verify projections and deliverables outside the processing engine.

Need the whole ecosystem? Open the drone work software guide.

1. Step 1 of 5

   Learn coordinate systems before buying RTK

   Understand raster vs. vector, latitude/longitude vs. projected coordinates, horizontal and vertical datums, units, GeoTIFFs, LAZ, and GeoPackage in the QGIS training manual.

2. Step 2 of 5

   Process one public dataset three ways

   Run the same imagery through a local open tool and a commercial trial. Compare alignment, holes, GSD, surface artifacts, reports, and export options.

3. Step 3 of 5

   Repeat one small site

   Fly the same legal, noncritical area several times. Keep height, overlap, camera angle, light window, and route consistent. Learn what changes break comparison.

4. Step 4 of 5

   Add independent checkpoints

   Partner with a surveyor or rent appropriate GNSS equipment. Keep some surveyed points out of adjustment, calculate error, and write an honest accuracy report.

5. Step 5 of 5

   Sell the narrow output

   Start with visual progress, inventories, or capture support. Do not market legal boundaries or certified topography until the licensed-professional and state-law questions are resolved.

These links are the starting point for current rules, methods, and professional boundaries. Vendor documentation explains a product; it does not replace the FAA, a regulator, a project specification, or a qualified reviewer.

• USGS: What is lidar data?

  LiDAR concepts and formal deliverable context.

• Esri Drone2Map control guidance

  Control points, checkpoints, and adjustment workflow.

• QGIS Training Manual

  Free GIS curriculum for projections, raster, vector, and analysis.

• NCEES Model Law

  A model definition of surveying; actual state law controls.