- Geography 5223 -
Project 3: Georeferencing Raster Images
Registering an Image for the Purpose of Data Creation

Objective:
georeference a raster image for the purposes of heads-up digitizing

The above two screen captures show the roads layer of State College imposed over an unrectified aerial photograph of the town of State College, PA. The photograph in the top screen shot contains no coordinate system and therefore the roads and their intersections of the two layers do not coincide. In order for the two layers to match up properly with the same coordinates, the layers must be georeferenced. Initially, the transformation of specifically marked control points on the photograph took place. Transformation signifies that the control points on the photograph will correlate to the same location on the roads layer. The roads layer already contains a coordinate system and therefore the points on the roads layer have real-world coordinates (the x-y coordinates defined by a map projection). After transformation has taken place, the five specified control points will have been placed within a geographical context. The transformation will also be applied to other factors in the output data set. The process of warping uses mathematics and statistics in order to make data as accurate as possible while creating correct relationships between objects in the image and those on the earth. This is also considered geometric transformation, but involving a higher order of polynomial equations and statistics.

The bottom capture shown above demonstrates the two layers which are georeferenced. Both the roads layer and the aerial photograph have the same coordinate system and the intersections on the roads layer match up with those on the aerial photograph.

The table below shows the root-mean-squared (RMS) error after georeferencing the five control points.