Automating the Geo-Referencing of Historic Aerial Photography in Flanders

In the Flanders region of Belgium, a vast archive of historical aerial photographs continues to be digitized, with new images regularly needing to be added to a public-facing digital platform. However, there’s a significant challenge: None of these photos are georeferenced, making them incompatible with the tool, which only supports georeferenced files. Traditionally, all the photos were organized solely by flight route and georeferenced manually. Without this critical geometry metadata, manually georeferencing every image would be an extremely time-consuming and labor-intensive process. To solve this, we developed an automated solution using FME to bridge the gap. By interpolating between a few georeferenced images, the process estimates the entire flight path. Using this we can then georeference all aerial photos along that path using the four corner points of each image based on the interpolated data with an acceptable degree of accuracy. While the process may not deliver centimeter-level precision, overlap between images ensures users can always view their areas of interest on the platform. The result? A dramatic reduction in manual effort, cost savings, and faster access to georeferenced images for the public. This case demonstrates how FME can significantly optimize large-scale data processes, delivering value through automation and efficiency.