hmmm… if that method doesn’t work, you might be able to set up the exclusions in a collection project, and then export by filtering on that project.
(not applicable in this case, but it might be worth knowing for the future: there’s a similar concept where you can filter by multiple taxa from a list, but i don’t think there’s an exclusion version of this filter.)
Instead of a single output field image_url, I would suggest that the following would be very useful if available in the exported CSV: image_url_stem – the basal part of the url except for the final slug (e.g., excluding image1.jpeg) number_of_images: how many photos image_filenames: a list of file names separated by a pipe operator for instance (as is recommended for GBIF), e.g., image1.jpeg | image2.jpeg
For analysis or downloads of all photos, image_url_stem can be concatenated with disaggregated image_filenames easily.
This will be a very useful feature to add, I feel. Thanks.