Here my YOLOv11Segmentation2GeoJSON script which allows you to create GeoJSON files (for use in GIS environments) from the output (prediction) of YOLO v11 Segmentation models.
When to use this script?
For example if you are dealing with (georeferenced) images taken by drone in which there are objects that you want to annotate and at the same time you want to retain the georeferencing information. If you process a TIFF (or GeoTiff ) or other format image on YOLO, the georeferencing information is lost.
With this script, on the other hand, you will obtain in output a GeoTiff file with the annotations made by the YOLO model (you can of course also use a YOLO model trained by you, just indicate in the script the name of the file that contains the model) as well as a GeoJson file that can be used in a GIS environment (e.g. QGIS). The GeoJSON file will obviously also contain the georeferencing information of the various objects