Orthorectification Simulator: Correcting Aerial Image Geometry

simulator intermediate ~10 min
Loading simulation...
d = 5.0 m — relief displacement at image edge

At 500 m altitude with 50 m terrain relief, objects at the image edge are displaced by approximately 5 m from their true ortho position — requiring DEM-based correction for mapping.

Formula

d = Δh × r / H (relief displacement)
GSD = H × pixel_size / f (ground sample distance)
S = f / (H - h) (image scale at elevation h)

From Perspective to Plan View

An aerial photograph is a perspective projection — objects near the camera appear larger, and elevated terrain features lean radially away from the image center. Orthorectification transforms this perspective view into a planimetric (map-like) projection where every pixel corresponds to a consistent ground distance. The corrected image, called an orthophoto, can be overlaid directly on maps and used for accurate area and distance measurements.

Relief Displacement

The most significant distortion in aerial imagery is relief displacement: elevated objects appear shifted outward from the image center by an amount proportional to their height and radial distance. A building 50 m tall photographed from 500 m altitude at the image edge is displaced by approximately 5 m from its true planimetric position. Correcting this requires a digital elevation model that specifies the terrain height at every point.

The DEM Connection

Orthorectification accuracy depends critically on DEM quality. A DEM error of 1 m at 500 m flight altitude introduces approximately 0.2% positional error — acceptable for many applications but significant for precision mapping. LiDAR-derived DEMs with decimeter accuracy enable orthophotos that meet the strictest cartographic standards. Photogrammetric DEMs generated from the same image set create a bootstrapping challenge that iterative refinement addresses.

Orthomosaics at Scale

Modern drone mapping stitches hundreds of orthorectified images into seamless orthomosaics covering entire construction sites, farms, or mining operations. Color balancing, seamline optimization, and ghosting removal at image boundaries ensure visually consistent results. The combination of centimeter GSD with rigorous geometric correction makes drone orthomosaics a standard tool in surveying, agriculture, and infrastructure inspection.

FAQ

What is orthorectification?

Orthorectification removes geometric distortions from aerial or satellite images caused by camera tilt, lens distortion, and terrain relief. The result is an orthophoto — a geometrically correct image where every pixel represents a consistent ground distance, suitable for mapping and GIS measurement.

Why do aerial photos need orthorectification?

Aerial photographs are central projections — objects closer to the camera appear larger, and elevated features lean outward from the image center (relief displacement). Without correction, distances and areas measured from raw photos are systematically wrong, with errors proportional to terrain relief.

What data is needed for orthorectification?

Orthorectification requires camera parameters (focal length, principal point, distortion), exterior orientation (camera position and attitude), and a digital elevation model (DEM). Ground control points (GCPs) may be used to refine accuracy. The DEM is critical — it corrects relief displacement.

What is ground sample distance (GSD)?

GSD is the ground area represented by one image pixel, typically in centimeters per pixel. It depends on flight altitude, focal length, and sensor pixel size: GSD = H × pixel_size / f. Smaller GSD means higher spatial resolution. Drone surveys commonly achieve 1-5 cm GSD.

Sources

Embed

<iframe src="https://homo-deus.com/lab/photogrammetry/orthorectification/embed" width="100%" height="400" frameborder="0"></iframe>
View source on GitHub