What is photogrammetry?

TL;DR

Photogrammetry is the science of creating 3D measurements and models from photographs. By taking overlapping images of an object or site from multiple angles and using software to identify matching points across those images, photogrammetry calculates the 3D position of every visible surface point. A drone photogrammetry survey can capture a 50-hectare site in under an hour and produce a 3D model accurate to 2-5 centimetres. Photogrammetry is used across mining, construction, and manufacturing for topographic mapping, stockpile volumes, progress monitoring, and as-built documentation. It is cheaper than LiDAR for many applications but cannot penetrate vegetation and requires good lighting conditions.

Key Takeaways

• Photogrammetry creates 3D coordinates from 2D photographs using a technique called "structure from motion" combined with "multi-view stereo"
• Drone photogrammetry with RTK positioning achieves 2-5 cm horizontal accuracy and 3-8 cm vertical accuracy—sufficient for most topographic and volumetric applications
• A typical drone photogrammetry flight captures 200-1,000 images, which are processed over 4-12 hours to produce a 3D model, orthomosaic, and digital elevation model
• Photogrammetry requires good lighting, clear visibility, and sufficient image overlap (typically 70-80% front overlap, 60-70% side overlap)
• Unlike LiDAR, photogrammetry cannot penetrate vegetation canopy or operate effectively in low-light conditions

Table of Contents

• What is photogrammetry?
  • Definition: photogrammetry explained
  • How photogrammetry works: the process step by step
  • Types of photogrammetry: aerial, terrestrial, and close-range
  • Accuracy: how precise is photogrammetry?
  • Photogrammetry software: processing and outputs
  • Photogrammetry vs LiDAR: a direct comparison
  • Drone photogrammetry: equipment and workflow
  • Applications in mining, construction, and surveying
  • Photogrammetry costs: 2025 Australian rates
  • Frequently asked questions
  • What to do next

Definition: photogrammetry explained

Definition: Photogrammetry is the science, technology, and art of obtaining reliable spatial measurements and three-dimensional information about physical objects and environments through the process of recording, measuring, and interpreting photographic images. It uses the geometric principles of perspective and triangulation to derive 3D coordinates from 2D image observations.

In practical terms: you take many photographs of something from different angles, and software figures out where everything is in 3D space. The result is a dense point cloud, a textured 3D mesh, an orthomosaic (a geometrically corrected aerial image), and a digital elevation model.

Photogrammetry has been used in surveying since the mid-19th century, but the technology transformed with the advent of digital photography, drone platforms, and automated processing software. What once required specialised cameras, darkrooms, and manual plotting now happens automatically in cloud-based or desktop software.

How photogrammetry works: the process step by step

Step 1: Image acquisition

The photogrammetric process begins with capturing overlapping photographs. The overlap is critical—without it, the software cannot identify matching points.

Step 2: Camera calibration and interior orientation

The software uses the camera's known parameters—focal length, sensor size, lens distortion characteristics—to establish the mathematical relationship between image coordinates and real-world directions. This is called "interior orientation."

Step 3: Aerial triangulation (structure from motion)

The software analyses all images simultaneously, identifying thousands of matching features across overlapping photos. Using a technique called "structure from motion" (SfM), it calculates:

• The 3D position of every matched feature point
• The precise position and orientation (X, Y, Z, omega, phi, kappa) of each camera when each photo was taken

This process requires no prior knowledge of camera positions. The software solves everything simultaneously in a process called "bundle adjustment."

Step 4: Georeferencing

The model is scaled and positioned in a real-world coordinate system using ground control points (GCPs)—targets with known coordinates surveyed independently with a GNSS receiver or total station. Alternatively, RTK/PPK drone positioning provides direct georeferencing without GCPs.

Step 5: Dense matching (multi-view stereo)

Once camera positions are known, the software performs "dense matching"—calculating a 3D position for virtually every pixel in every image. This creates a dense point cloud with millions or billions of points.

Step 6: Output generation

From the dense point cloud, the software generates:

Types of photogrammetry: aerial, terrestrial, and close-range

Drone photogrammetry

The most common type for industrial applications in Australia. A drone carries a high-resolution camera and flies a programmed grid pattern over the survey area. Modern RTK-equipped drones (DJI Matrice 350 RTK with Zenmuse P1) achieve 2-3 cm accuracy without extensive ground control.

Terrestrial photogrammetry

Uses cameras mounted on tripods or handheld rigs to capture buildings, structures, and objects from ground level. Often combined with laser scanning to add photographic texture (RGB colour) to point cloud data. Terrestrial photogrammetry can achieve sub-millimetre accuracy at close range.

Close-range photogrammetry

Used for small objects and precision measurement. Industrial applications include quality control of manufactured parts, forensic documentation, and heritage recording. With calibrated cameras and coded targets, close-range photogrammetry achieves 0.01 mm accuracy.

Accuracy: how precise is photogrammetry?

Photogrammetry accuracy depends on camera quality, flying height, overlap, ground control, and processing quality.

The rule of thumb for drone photogrammetry: horizontal accuracy is approximately 1-2 times the GSD; vertical accuracy is approximately 2-3 times the GSD. So at 2 cm GSD, expect 2-4 cm horizontal and 4-6 cm vertical accuracy with good ground control.

Photogrammetry software: processing and outputs

Pix4Dmapper and Agisoft Metashape are the two most widely used packages for industrial drone photogrammetry in Australia. Pix4D has stronger cloud processing and project management tools. Metashape offers a perpetual licence option and more flexibility for non-standard workflows.

Photogrammetry vs LiDAR: a direct comparison

Key point: The choice between photogrammetry and LiDAR comes down to three questions: (1) Does your site have vegetation? (2) What are your lighting conditions? (3) Is visual texture important for your deliverable? If the answer to question 1 is yes, choose LiDAR. If questions 2 and 3 favour photogrammetry and the site is clear, photogrammetry is likely more cost-effective.

Drone photogrammetry: equipment and workflow

Typical equipment setup

Standard workflow

1. Flight planning: Define survey area, set overlap (75% front, 65% side), calculate flight lines, set altitude for target GSD
2. GCP placement: Survey 5-10 ground control points with GNSS RTK or total station; distribute across the site
3. Flight execution: Launch drone, monitor progress, verify image quality during flight
4. Data download: Transfer 200-1,000 images to processing workstation
5. Processing: Import to photogrammetry software, run aerial triangulation, review quality report
6. Dense matching: Generate dense point cloud and DEM
7. Output generation: Create orthomosaic, contours, and any derived products
8. Quality control: Check accuracy against independent check points; verify coverage
9. Delivery: Provide orthomosaic, DEM, point cloud, and report in specified formats

Applications in mining, construction, and surveying

Mining

Construction

Surveying

Photogrammetry costs: 2025 Australian rates

Photogrammetry is consistently 20-40% cheaper than LiDAR for equivalent site coverage on clear terrain. The cost advantage diminishes on vegetated sites where LiDAR's ground-penetration capability saves significant field and processing time.

Frequently asked questions

How many photos does a photogrammetry survey need?

A typical drone photogrammetry project captures 200-1,000 images for a 10-50 hectare site. The exact number depends on the survey area, overlap settings, camera sensor size, and flight altitude. Processing software provides estimates during flight planning.

Can photogrammetry be used for measuring stockpile volumes?

Yes. Photogrammetry-derived DEMs are widely used for stockpile volume calculations. Accuracy is typically 3-5% of true volume when ground control is used. For high-value materials or reconciliation purposes, LiDAR may be preferred for its slightly higher accuracy and consistency.

What weather conditions prevent drone photogrammetry?

Winds above 35-40 km/h, rain, and heavy cloud cover can prevent or degrade photogrammetry flights. Rain damages equipment and degrades image quality. Strong winds affect drone stability and image sharpness. Heavy cloud reduces light, increasing noise and requiring longer exposures that cause motion blur.

How does photogrammetry compare to traditional ground survey for topography?

For large areas (over 10 hectares), drone photogrammetry is typically 40-60% cheaper and 5-10x faster than traditional total station or GPS survey. For small, detailed sites (under 2 hectares), the cost difference narrows and ground survey may be more accurate for specific engineering features.

What is the difference between an orthomosaic and a normal aerial photo?

An orthomosaic is geometrically corrected so that every pixel is shown as if viewed directly from above (orthorectified), with scale consistent across the entire image. A normal aerial photo has perspective distortion—objects appear to lean away from the camera and scale varies with distance. An orthomosaic can be used for accurate measurement; a normal photo cannot.

What to do next

Photogrammetry is the most cost-effective way to capture large-area 3D data for clear sites. If your project involves topographic mapping, progress monitoring, or volume calculation on a site without dense vegetation, it should be your first consideration.

1. Assess your site conditions. Is the terrain mostly clear? Is vegetation minimal? If yes, photogrammetry is suitable. If dense vegetation covers the ground, consider LiDAR instead.
2. Define your accuracy requirement. Standard engineering topographic work needs 3-5 cm accuracy, which is achievable with RTK drone photogrammetry. Higher precision may require additional ground control or terrestrial methods.
3. Plan for processing time. Field capture is fast—often under an hour for a typical site. Processing takes 4-12 hours. Rush jobs cost more.

Call ISS on 0407 057 015 to discuss your photogrammetry requirements. We operate DJI RTK drone systems with Zenmuse P1 cameras and process in Pix4Dmapper and Agisoft Metashape. We will assess your site, accuracy needs, and budget and recommend the right capture and processing approach.