Could it be possible to create a transformer that uses AI image recognition to georeference for example aerial images? AI detects similarities between image and reference imagery and creates reference points that can be used in georeferencing.
Page 1 / 1
My colleagues have made a couple of interesting transfomers on FME Hub like the OpenAIVisionConnector. These transformers take images and can answer questions about said images. It does not quite do what you describe, but if questions on images can be answered, I’m sure your case is possible to create as well.
Check out this linkedin post for more info - https://www.linkedin.com/posts/oliver--morris_tensing-fme-fmepartners-activity-7237775256436764673-Z0UY?utm_source=share&utm_medium=member_desktop
New→Open