How to Create Orthomosaics from Drone Photos: Step-by-Step
Turn hundreds of drone photos into one seamless, distortion-free map. Here's the complete orthomosaic creation workflow.
Creating an orthomosaic involves turning drone photos into a geometrically corrected, distortion-free 2D map. The process requires quality data capture with high overlap and ground control points, followed by a six-step software workflow: importing photos, aligning features, building a dense point cloud, generating a 3D mesh, projecting the orthomosaic, and exporting as GeoTIFF.
An orthomosaic is more than a stitched panorama. A panorama blends images together with warped buildings and distorted distances. An orthomosaic is geometrically corrected — optical distortion, terrain distortion, and camera tilt are all removed. Every pixel is assigned a real-world geographic coordinate.
Before You Process: Capture Quality Data
- Overlap: 75% front, 65% side minimum. More overlap = better accuracy.
- Consistent altitude: Steady flying height. Use terrain following if available.
- Good lighting: Overcast days are ideal — clouds diffuse light and reduce harsh shadows.
- GCPs: Non-negotiable for survey-grade accuracy. Visible targets surveyed with RTK GPS.
The Step-by-Step Workflow
Step 1: Import Photos
Drag and drop your folder of JPEGs into your photogrammetry software. The software reads EXIF metadata — GPS coordinates, camera sensor dimensions, focal length, and lens distortion parameters. Verify all images loaded and the camera model was detected correctly.
Step 2: Align Photos (Feature Matching)
The software analyzes every image to find matching visual features — building corners, rocks, pavement changes — across overlapping photos. Once “tie points” are identified, geometry calculates the exact 3D position and orientation of the camera for every shot.
Output: a sparse point cloud and camera alignment visualization. Poor results here usually mean insufficient overlap, blurry images, or featureless terrain.
Step 3: Build Dense Point Cloud
The sparse cloud is a basic framework. Building a dense cloud goes further — calculating depth for every pixel in every image. The result is millions or billions of 3D points representing the true shape and topography of the terrain.
Step 4: Generate 3D Mesh
The software connects dense cloud points into a continuous surface of triangles (TIN — Triangulated Irregular Network). Think of stretching tight digital fabric over the point cloud. For top-down mapping, a height-field mesh is most efficient.
Step 5: Create Orthomosaic Projection
The software projects original 2D photos onto the 3D mesh. Because the mesh accounts for terrain elevation and camera tilt, all distortions are corrected. The result: one seamless, geometrically accurate, uniform-resolution 2D map.
Step 6: Export GeoTIFF
Export as GeoTIFF — the industry standard for geographic data. Unlike JPEG/PNG, GeoTIFF embeds spatial metadata including the Coordinate Reference System (CRS). When opened in QGIS or ArcGIS, it drops into the correct location automatically. Use LZW compression to reduce file size without quality loss.
Recommended Processing Settings
| Step | Parameter | Setting | Notes |
|---|---|---|---|
| Align Photos | Key Point Limit | 40,000 | Balances speed and detail |
| Align Photos | Tie Point Limit | 4,000 | Keeps sparse cloud clean |
| Dense Cloud | Quality | Medium | Use High for small sites |
| Dense Cloud | Depth Filtering | Moderate | Removes noise from moving objects |
| 3D Mesh | Surface Type | Height Field | Best for top-down maps |
| Orthomosaic | Blending Mode | Mosaic | Prevents ghosting |
| Export | Format | GeoTIFF | Standard for GIS |
| Export | Compression | LZW | Lossless, smaller files |
Frequently Asked Questions
How does an orthomosaic differ from a standard panorama? Unlike a panorama that warps buildings and distorts distances, an orthomosaic is geometrically corrected, removing optical and terrain distortions with a precise real-world coordinate for every pixel.
What conditions are needed for quality data capture? At least 75% front and 65% side overlap, consistent flying altitude, overcast lighting to minimize shadows, and RTK-surveyed ground control points for survey-grade accuracy.
What happens during the photo alignment phase? The software identifies matching visual features across overlapping images to establish tie points, calculating the exact 3D position and orientation of the camera for every captured shot.
Why is a 3D mesh required for the final map? The mesh represents the true shape and topography of the terrain. Projecting original photos onto this 3D mesh eliminates camera tilt and elevation distortions.
Why export as GeoTIFF? GeoTIFF embeds spatial metadata including the Coordinate Reference System, ensuring the map automatically drops into the correct location when opened in GIS software like QGIS.
What file compression is recommended? LZW compression reduces file size without any loss of image quality. This lossless approach keeps your final map sharp while saving storage space. Master the complete mapping pipeline in our free Drone Mapping course — from flight planning to final export.
