It would be nice to be able to use the “computer vision” without having to create an observation. A common process in the field for me is:
- I see a plant that I don’t know at all
- Take a few pictures with my phone, and try them in the computer vision to see what comes up.
- Use those suggestions to inform my identification (e.g. often the suggestions will give me the wrong species but in the right genus or family, and I can then use a key to identify to species)
- Use my proper camera to take photos that show all the necessary identification features, to upload to iNat later.
This is amazing. I can often quickly identify things that aren’t flowering, which is often nearly impossible otherwise. And even for things with flowers, I can often get a confident species identification in a minute or two, rather than taking forever to go through a giant key or flip through a field guide.
Maybe this is a super niche use. If so, it really shouldn’t be - I think this is now by far the best way to learn your plants as a beginner to intermediate level botanist. I rarely touch my plant field guides anymore except to browse at home.
Step two is currently pretty annoying (I have to create a new observation, and then switch the “primary photo” every time I want to try a new photo). And sometimes I’ll accidentally create useless observations with a bunch of photos of different species. It would be nice to just have a “computer vision” button to test any photo.