Aerial photogrammetry — the science of extracting precise measurements and 3D geometry from overlapping aerial photographs — has been a cornerstone of surveying and mapping for over a century. The arrival of consumer and commercial drones transformed aerial photogrammetry from a specialty discipline requiring manned aircraft into an everyday field tool. This guide explains what aerial photogrammetry is, how the drone-based workflow operates, where accuracy comes from, which industries rely on it, and how videogrammetry (a newer approach pioneered by platforms like SkyeBrowse) fits into the picture.
Key Takeaways
- Aerial photogrammetry derives precise measurements and 3D geometry from overlapping drone photographs
- With ground control points, horizontal accuracy of 1-2 cm and vertical accuracy of 2-5 cm is achievable
- Videogrammetry (video-based capture) reduces field time from 20-40 minutes to 3-5 minutes per site
- Used across surveying, construction, public safety, insurance, agriculture, and heritage preservation
Contents
- What is aerial photogrammetry?
- How does the drone photogrammetry workflow work?
- What determines accuracy in aerial photogrammetry?
- Which industries use aerial photogrammetry?
- How does videogrammetry compare to photo-based aerial photogrammetry?
- FAQ
What is aerial photogrammetry?
Aerial photogrammetry is the technique of deriving measurements, maps, and 3D models from photographs taken from above — traditionally from aircraft, now routinely from drones. Overlapping images taken at known positions allow software to triangulate the 3D coordinates of every visible point on the ground, producing outputs such as orthomosaics (geometrically corrected aerial maps), digital elevation models (DEMs), and textured 3D meshes.
The word "photogrammetry" combines the Greek roots for light (photo), writing (gram), and measurement (metry) — and the field has been codified in professional standards by organizations such as the American Society for Photogrammetry and Remote Sensing (ASPRS) since the early twentieth century. Traditional aerial photogrammetry used large-format metric cameras aboard fixed-wing aircraft to map entire regions. Drone photogrammetry retains the same geometric principles at a fraction of the cost, bringing sub-meter and even centimeter-level mapping to small sites, individual structures, and time-sensitive scenes.
The core mathematical engine is Structure from Motion (SfM), which identifies matching features across hundreds of overlapping images and back-calculates camera positions and a dense 3D point cloud simultaneously. The outputs of that point cloud — orthomosaics, DEMs, and mesh models — form the deliverables that surveyors, engineers, and public safety teams act on. See the drone mapping guide for a broader treatment of flight planning and deliverable types.
How does the drone photogrammetry workflow work?
A drone photogrammetry mission has four phases: flight planning, data capture, processing, and delivery. During data capture, the drone flies a grid or orbital pattern and fires its camera at regular intervals, producing images with 70-90% forward overlap and 60-80% side overlap. Processing software then aligns the images, builds a dense point cloud, and generates the desired map or model outputs.
Flight planning sets altitude, overlap, and flight path. Higher altitude means faster coverage but coarser resolution (larger ground sampling distance, or GSD). A GSD of 2-3 cm per pixel is typical for construction or survey work; crash scene documentation often targets 1 cm or finer, requiring lower altitude or a higher-resolution sensor.
Data capture is performed autonomously by mission-planning apps (DJI Pilot 2, DroneDeploy, Pix4Dcapture, and others). The drone records GPS position and attitude for every image trigger, embedding this metadata as EXIF/telemetry data that anchors the photogrammetric reconstruction.
Processing runs in desktop software (Agisoft Metashape, Pix4Dmapper, RealityCapture) or in the cloud. The pipeline follows a standard sequence: feature detection and matching, sparse point cloud (Structure from Motion), dense point cloud reconstruction (Multi-View Stereo), mesh generation, texture mapping, and orthomosaic export. Processing a 400-image survey at high quality can take 1-4 hours on a GPU-equipped workstation.
Delivery formats include GeoTIFF orthomosaics, LAS/LAZ point clouds, OBJ or GLB meshes, and shapefiles — all geo-referenced to real-world coordinates if GCPs were used.
What determines accuracy in aerial photogrammetry?
Accuracy in aerial photogrammetry is governed by three main variables: ground sampling distance (image resolution at ground level), image overlap percentage, and the use of ground control points (GCPs). Without GCPs, relative accuracy is typically 1-3 times the GSD; with surveyed GCPs, absolute accuracy can reach 1-2 cm horizontally and 2-5 cm vertically on well-textured surfaces.
Ground control points are precisely surveyed markers placed in the scene before flight. Their coordinates are measured with a total station or RTK GNSS receiver and then matched to their pixel positions in the imagery, allowing the photogrammetric reconstruction to be anchored to a real-world datum. The USGS and ASPRS both publish accuracy standards that specify minimum GCP count and distribution for mapping projects at various quality levels.
When GCPs are impractical — emergency scenes, hazardous areas, time-critical incidents — RTK/PPK drones can geolocate each image to 1-3 cm accuracy using an onboard GNSS receiver that logs corrections from a base station. This allows high accuracy without placing ground markers. For detailed guidance on GCP setup and when to skip them, see the ground control points guide.
Overlap percentage is the other critical lever. The IEEE and academic literature consistently show that dropping front overlap below 70% causes tie-point gaps and reconstruction artifacts, particularly on flat or uniform terrain where feature matching is harder. More overlap means more processing time but fewer failures.
Which industries use aerial photogrammetry?
Aerial photogrammetry is used across surveying and civil engineering, public safety, construction, insurance, agriculture, environmental monitoring, and cultural heritage preservation. Each industry leverages the same core outputs — orthomosaics, DEMs, and 3D meshes — for different decision-making workflows, from site volumetric calculations to crash scene documentation.
Surveying and civil engineering: Topographic surveys, corridor mapping for roads and pipelines, earthwork volume calculations, and boundary confirmation. Drone photogrammetry can complete in hours what a ground survey crew takes days to measure.
Public safety: Law enforcement and fire departments use drone photogrammetry to document crash scenes, crime scenes, structure fires, and disaster areas. The 3D model provides a permanent, court-admissible record that can be revisited without revisiting the scene. Learn more about how drones serve fire operations at the fire department operations guide (internally linked via the broader site).
Construction: Progress monitoring, as-built verification, cut-and-fill calculations, and coordination with BIM models. Repeat drone flights over the same site at weekly intervals create a time-series of the project's advancement.
Insurance: Roof assessments, catastrophe claims, and property documentation after storms or wildfires. Aerial photogrammetry gives adjusters a defensible, geo-referenced record without the safety risks of physical roof access.
Agriculture: Crop health monitoring using multispectral cameras alongside RGB photogrammetry, stand-count mapping, and irrigation planning.
Archaeology and cultural heritage: High-resolution orthomosaics and 3D models of excavation sites, historic structures, and artifacts, allowing non-destructive documentation at millimeter scale.
How does videogrammetry compare to photo-based aerial photogrammetry?
Videogrammetry applies the same photogrammetric reconstruction algorithms to continuous video frames rather than pre-planned still images. Because video captures hundreds of frames during a single freehand orbital flight, the operator does not need to plan a grid, set overlap parameters, or execute a structured mission — they simply fly around the subject. This reduces field time from 20-40 minutes per site to 3-5 minutes while still producing georeferenced 3D models, orthomosaics, and point clouds.
Traditional aerial photogrammetry with still images has two practical limitations: mission planning takes time, and the grid flight pattern is rigid. If something moves, the light changes, or an emergency requires fast action, the structured photo-capture approach can become a bottleneck. Videogrammetry removes the grid entirely — a pilot flies a short orbital video, uploads the footage, and the cloud processes it into a full 3D deliverable.
SkyeBrowse is a cloud-based videogrammetry platform used by more than 1,200 agencies worldwide, including law enforcement, fire departments, and construction teams. It accepts .MP4 and .MOV video via browser upload or the SkyeBrowse Flight App, ingests DJI (.SRT) and Autel (.ASS) telemetry for georeferencing, and delivers LAZ point clouds, GLB 3D meshes, and GeoTIFF orthomosaics — the same output formats as photo-based photogrammetry software. Accuracy tiers range from Lite (2-6 inch) to Premium Advanced (0.1 inch with 16K processing and AI moving-object removal).
The tradeoff versus traditional aerial photogrammetry software is coverage area per flight: grid photogrammetry can map hundreds of acres in a single mission, while videogrammetry excels at individual structures, intersections, and scenes up to a few acres. For large-area survey work — agricultural fields, corridor mapping, utility rights-of-way — photo-based drone photogrammetry remains the standard. For rapid situational mapping of a specific site, videogrammetry is consistently faster.
For a detailed head-to-head comparison of the two approaches, including accuracy benchmarks and workflow diagrams, see the videogrammetry vs. photogrammetry comparison. If you are choosing between photogrammetry and LiDAR for your project, the LiDAR vs. photogrammetry guide covers cost, accuracy, and use-case fit in depth.
FAQ
What is the difference between aerial photogrammetry and LiDAR?
Aerial photogrammetry reconstructs 3D geometry from overlapping photographs using computer vision algorithms, while LiDAR uses pulsed laser beams to measure distances directly. Photogrammetry produces photo-realistic textured models but requires adequate lighting and surface texture; LiDAR penetrates canopy and works in low light but costs significantly more per flight. See the full LiDAR vs. photogrammetry comparison for a project-by-project decision guide.
How accurate is drone photogrammetry?
Without ground control points, relative accuracy is typically 1-3 times the ground sampling distance (GSD). With surveyed GCPs distributed across the survey area, absolute horizontal accuracy of 1-2 cm and vertical accuracy of 2-5 cm is achievable on well-textured surfaces at low altitude. RTK/PPK drones can approach similar accuracy without physical ground markers. The ground control points guide covers GCP placement strategy in detail.
What is videogrammetry and how does it differ from photogrammetry?
Videogrammetry applies photogrammetric principles to continuous video frames rather than discrete still photographs. Because a single orbital video flight captures hundreds of overlapping frames, there is no need to plan a grid or trigger a camera at intervals — the pilot simply orbits the subject and uploads the file. Platforms like SkyeBrowse use this approach to generate 3D models, orthomosaics, and point clouds in minutes from footage shot with off-the-shelf DJI or Autel drones, at a fraction of the field time required by traditional photo-based aerial photogrammetry software.
