Moderately often (I mainly ID birds), I see observations with multiple, at times 10 or more, very similar pictures.
For my bird observations I only add the best one or two pictures, sufficient to give enough data to get a ID confirmation. For a plant, it would be the whole plant, leaves, flower, bud, fruit etc (if possible). For a butterfly I might try to get a view with wings open and closed to see both sides.
Not I get that many pictures (including poor ones) help train the AI and if poor shots a few extra can help (if a different angle or different part obscured by a stick) but surely there is a limit?
When confirming an ID it does imply you need to look at all pictures in case one is of another species (It does happen for birds as they form mixed flocks of related species). At times I’ll admit I do not do that
I don’t think I’ve ever done as many as 10, but I often upload what are probably redundant pictures of invertebrates because I know their ID often hinges on little things like how many segments the antennae have. But for most of them, I don’t know which tiny differences are important for that order/genus so I just try to get as many clear angles as possible.
Well, I suspect it depends a bit on what is being observed…
I observe and ID a lot of bees – here it can absolutely be helpful to include multiple photos even if they superficially look similar because sometimes a subtle trait like, say, the color/shape of the spurs on the hind legs can be be crucial for determining the species.
This doesn’t necessarily mean that one has to upload dozens of photos. Even for difficult taxa, I suspect that between 3-6 photos are probably absolutely sufficient in most cases – if they are well chosen to show the relevant features. The problem is that the observer may not know what those relevant traits are and which of their photos are most useful for IDers, so they might choose to upload everything to increase the chances of getting a precise ID.
Sometimes I think a desire to share one’s photos of what one witnessed may also play a role. I’m not a birder and always excited if I manage to get halfway decent bird photos. So in such cases I may very well upload a few more photos than are strictly necessary, though I usually try not to upload multiple similar photos.
The other time I might upload extra photos is to document some behavior or interaction (a spider building a web, a wasp carrying prey to a nest, etc.).
I agree it depends on what you’re observing. As has already been mentioned, for arthropods where every little detail counts, even photos apparently very similar can provide the key to distinguishing not just the species, but sometimes even the genus. The same can also be true for plants where two similar photos can sometimes reveal the presence/absence of hairs, the underside of a leaf, etc. Ten does seem rather a lot, perhaps 5/6 should be the maximum, but every observation is different.
I upload single or multiple pictures on what I know / don’t know , and also what I suspect are easier or harder to identify. Or if I can (or have) positively id something but the picture is not great in the hope that it may make the next algorithm better.
For things I don’t know I usually post multiple angles and closes up of different parts (if feasible) as @ sarahtheentwife says.
For plants and trees usually different parts again - leaves upper and under, bark / stem if relevant, leaves from afar so the formation is seen, and then flowers closeup and far. For trees i try and take the bark as they may be relevant for winter identifications. Fruit at different stages if feasible (of course separate observations. Same thing with leaf, flower and fruit buds.
For nice subjects and things I think are rare I post more just so that others can enjoy them as much as I did (and again different angles might be rare)
Of late I have been trying to be efficient and take just one picture of things I know I have already seen .
The landscape and habitat - are the most difficult to incorporate into pictures - so sometimes I take the same subject with both a Dslr camera and a mobile phone camera (wider angle) and post that together.
Finally I am never satisfied with the pictures I have taken so end up repeating the same species in the home that I will get the perfect picture which some one may find good use for it.
I agree with what other users have written about the value of multiple shots and adding them as “insurance” when one doesn’t know the specific distinguishing characters that are needed.
In regards to this specific part of your question:
In general, the species the observation is for should be present in each photo in an observation - could be a different individual, could be anywhere in the pic, so for mixed flocks this could be tricky to tell. It’s ok if the pics have other species in them, as long as the focal species is somewhere in there. If it’s obvious that this isn’t the case for an observation (it’s a dump of a bunch of different species in a flock with each shot showing specific species that differ from other pics), I would ID to “Birds” or some other shared taxon you know of and leave a comment asking the user to split out pics to different observations and leave a comment when they’ve done so, so a more specific ID can be given on each. But if there’s a ton of pics, I usually just flip through quickly to make sure they make sense, and don’t scrutinize each one.
Interesting point about training the AI; I suspect, though, that there isn’t a lot to be gained from just adding tons of redundant photos, since this won’t make a difference where species (as opposed to, say, Genus) can’t be identified from a photo in any case.
Re. the larger question about the value of multiple photos: over time you will learn what the identifying features are for the creatures that interest you; you will learn that many species can’t be identified from a single shot; and that just taking many photos rather than 1 isn’t necessarily the answer, but instead the goal is to capture those identifying features (e.g. for a wasp, wing venation, etc.–try to get a good wing shot!)
Field guides are a wonderful place to begin because good ones will tell you what features to look for, and which species can be identified from photos (or a loupe).
For me, I have a few reasons I would upload more than, say, 3-5 images. The vast majority of my observations (vertebrates) are single images, and most are birds that don’t require so many different angles. Still, there are instances in which I upload several images even if one is technically sufficient for identification.
Sometimes, I’ve just captured multipleangles of the same specimen, and they are all too interesting not to share.
Occassionally, I’ve chased a rare bird and simply obtained a lot of great images, primarily to ensure that the reporting is accepted on eBird (which I also upload bird images to).
In my search for observations with many images, I found this observation of mine with 17 images that certainly doesn’t need so many of basically the same angle. I uploaded this observation during my initial mass upload when I was still new to iNat (and took the images when I was still new to my camera). Today, I might upload an image of it still/calling, an image of its perch adjustment, and an image of it taking off.
I use iNaturalist mostly for sharing images with the secondary benefits of filling in range gaps and getting identifications, so your reasons for adding multiple images/sounds (or not) will vary!
If small parts are needed for ID (insects, grasses, etc.), multiple similar photos can be useful. Not only is there the issue of a needed trait being shown, but is that trait in focus??? Too often not. Multiple photos increase the chance that at least one will have a given important trait in focus.
A different issue is that multiple parts or angles may be needed for ID (e.g. in many plants). And of course if I get really good and/or interesting photos I may post them though they’re not needed for ID.
iNaturalist puts a maximum of 20 on the number of photos we can upload per observation. I have met that barrier, though very, very rarely.
If I don’t know what something is, I try to use whatever pictures I ended up getting that are clear. Sometimes it’s many and sometimes it’s only a few. I never know what characteristic is needed so I just go for it. Maybe people won’t go through 10, maybe they will. There’s been times people have so you never know. I’d rather have more than I need than not enough.
This is something I often think of when using a macro lens to photograph insects or other small creatures. Becuase the depth of field is so thin, some of my shots have the face in focus, which is my goal, but in other photos the wings or other features are in focus. To me those are less aesthetically pleasing but may contain important diagnostic features like wing venation, so I try to include those as well.
As someone who mostly does macro photography, I support the idea that more angles are better. The problem is that it is very difficult for a non-expert like me to know which angles are needed. I try to compensate with multiple photos and focus stacking. Both have their pros and cons.
It depends whether or not iNat is simply about “confirming an ID”, or is it also about documenting behaviors, interactions, and ecological contexts? On the About Page of iNat it says this:
iNaturalist helps you identify the plants and animals around you while generating data for science and conservation. Get connected with a community of millions scientists and naturalists who can help you learn more about nature! What’s more, by recording and sharing your observations, you’ll create research-quality data for scientists working to better understand and protect nature. So if you like recording your findings from the outdoors, or if you just like learning about life, join us!
Admittedly, I upload a few more than I should at times, but who is to say who is going to be viewing your pictures? It might be a scientist from the other side of the planet who are seeking comprehensive data for their research, who knows? If they use them all well and good, if not that’s ok too. But they can’t use what’s not there to begin with.
I think the depth of field is narrow in some pictures. The uploader decides that more pictures offer more clues, like some pictures are sharp at the head of an insect, and another picture is focusing on the wings. Although the pictures can look the same in thumbnail size. Two or three pictures with a slightly different angle is a minor proof of authenticity. It implies that the uploader has a dozen photos of that observation in the computer. It is not known if AI can replicate that at this moment.
I think my only observations with more than 10 photos have been those where I am dissecting a gall to find out what is living inside. So I’m trying to show not only the plant gall structures, but also the structures of whatever may be living inside it. And because everything I’m looking at is relatively small, I’m trying to take macro photos with slightly different focuses in case it might show something significant.