georeference

Georeferencing is the process of assigning geographic coordinates to spatial data so that every location corresponds to a point on the Earth's surface. It links imagery, maps, or vector data to a common coordinate reference system, enabling integration with other spatial information and analysis in GIS.

The core components are a coordinate reference system (CRS), a datum, and a transformation. The CRS defines how coordinates relate to locations; the datum ties the CRS to the Earth's shape; a transformation resamples the data to fit the CRS, often using ground control points or sensor metadata.

Direct georeferencing uses sensor-provided position and orientation data at capture time (GPS/IMU) to assign coordinates. Indirect georeferencing relies on matching features in the image to known ground control points or features in a reference dataset, then computing a mathematical transformation.

Common transformations include affine (six parameters) for linear shifts, projective, and higher-order polynomial or rubber-sheeting for nonuniform distortion. Resampling methods include nearest neighbor, bilinear, and cubic convolution to create a new raster aligned to the target grid.

Applications include aligning satellite and drone imagery with maps, georeferencing historical images, cadastral maps, and environmental datasets for analysis in GIS.

Accuracy is assessed with check points and statistics such as RMSE. Limitations arise from uneven distribution of control points, sensor distortion, terrain relief, and changes over time. Metadata and standards, such as ISO 19115, describe the data and its georeferencing.

Georeferenced data are commonly stored in formats that embed coordinates or sidecar files, such as GeoTIFF with world files (TFW) or shapefiles with projected coordinates, and are managed in GIS platforms like QGIS and ArcGIS.