I have a bazillion or so (rough estimate) of wildlife photos dating back many years. I am trying to add them into iNaturalist, but have hit a bit of a snag in my workflow.
First I sort by date, then location. I then cull out bad or duplicate photos. I then upload them into Flickr albums. I then batch edit the location by placing them on a map. At this point they are ready to import (or are they?)
Most of these photos are not identified yet. I may have done so at some point in the past, but that information is not in any metadata or tag. I was hoping to do this as I import the observation into iNat, but the web site seems to be looking for me to already have the species ID in a tag field.
So this works great for already identified photos, but 95% of mine are not.
I could easily do this using the iOS app, but there is no connection between the app and Flickr. That connection only exists on the web site. And other copies of the photos that might exist, like in iCloud, donāt have any geotag.
So Iām stuck!
So what is the best way to get a large number of Flickr photos (mostly birds) into iNaturalist and get their identifications verified?
Help me understand the problem. Are you finding it too tedious to individually identify your photos? Are feeling that you donāt have enough knowledge to ID your photos with accuracy? Are you unfamiliar with the website, and you donāt know how to go about putting IDs on photos? Is uploading them without ID, and waiting for the community to do it, too slow? Strictly speaking, putting unidentified photos onto iNat is fine, even normal. For me, I find adding the location to old photos to be the painful part, so if you are past that, you seem to be doing well in my book.
To be more specific, the computer vision function does not appear to be available from the web interface. Thatās a major issue for a birding workflowā¦
The Flickr photo import page is one of the outdated pages on the website. The computer vision function is definitely available on the web interface in several other places, including the normal upload page. You can use computer vision on any observation by going to the individual observation page (e.g. https://www.inaturalist.org/observations/31804464), then where it says āSuggest an Identificationā, click where it says āSpecies nameā, and the top auto-suggestions appear in a dropdown.
Then I can upload my observations without IDs and go through them more quickly to add IDs than I can via the upload page. In the cases I want to use computer vision, I click the Suggestions tab and change the Source to Visually similar.
Thank you so much for this info! It really helps me clarify things a lot.
Because I intend(ed) to upload all of my bird photos to Flickr any way, I tried to avoid duplicate effort by importing existing photos from Flickr into iNat. But since that page does not have the Vision ID, if I try to save an entry with no text in the species field I get an error (fail whale/shark whatever that is).
It looks like my best workflow might be to upload directly then use the controls there to batch edit the location (works really well, better than Flickr) and adjust dates if needed. Then I can run the Vision ID per photo.
My only gripe is the inability to āzoom inā on the item in question from the web interface. Iāll have to manually crop before uploading, but without getting rid of the original photo. Guess Iāll need to do specific feedback on that.
Iāll still have to do a separate upload to Flickr, so my iNat workflow will be duplicate out of necessity, but at least this way I can combine Vision ID with some batch editing of location. Always trying to find shortcuts. Itās a curse! lol
You can zoom in a little bit by clicking on the photo:
But you canāt zoom in easily all the way, which I agree is annoying and another reason why I add IDs through the Identify interface instead. To view the full size image from the web uploader, you can click the photo, then right click on the photo pop-up and select āOpen image in a new tabā or āCopy image locationā, then change ālargeā to āoriginalā in the URL
"