Image Rectification
Definition
Image rectification transforms an image so that it has uniform scale and geometry with respect to a chosen coordinate system or another image. In GIS this often means correcting aerial or satellite imagery for sensor tilt, relief displacement, and lens distortions so that the output can be treated like an accurate map. Rectification uses ground control points and a mathematical model such as projective or polynomial transformations, or more advanced rigorous sensor models. The quality of rectification is measured by residual errors at control points and by visual inspection of linear features like roads and shorelines. Pixel size varies with off nadir view angles and terrain. Reporting effective ground sampling distance across the scene helps analysts judge whether the image supports their measurements. Where tall buildings cause layover, note residual distortions explicitly.
Application
Orthophotos produced by rectification support cadastre updates, infrastructure design, and change detection. Photogrammetry pipelines rectify images before mosaicking and 3D reconstruction. In emergency response, quickly rectified drone photos provide actionable context when official basemaps are outdated. Conservation groups rectify historical aerials to trace wetland loss or dune migration.
FAQ
What is the difference between rectification and orthorectification?
Orthorectification accounts for terrain using a DEM and sensor geometry to remove relief displacement. Simple rectification aligns to ground points but may leave height related distortions.
How many control points are required?
It depends on the transformation model. Use the minimum plus redundancy, distributed across the scene. Evaluate RMS error and examine local residuals along edges and corners.
Can we rectify images without surveyed control?
Yes, using image to image registration against a trusted orthophoto or using automatically detected tie points. Accuracy will reflect the reference quality.
What metadata should accompany rectified imagery?
Sensor, date, transformation type, control sources, DEM used, residual error statistics, and the coordinate reference system. These details enable informed reuse.