However, I still wonder if the iNaturalist platform or community could find a better way of encouraging and educating users about taking more and better photos. I know that there are resources available(the video tutorial, forum posts) that can guide users on taking photos for better identification, but one has to actively go looking for those. I wonder if occasionally broadcasting short tips in the same way that iNat posts things to the Dashboard banner might be a more effective way of incrementally educating users.
So many of the plant observations I look at cannot be identified to species or genus because they only contain one photo, and that photo does not include identifying features. I know this is a not a new issue, but I donāt see this situation changing unless we can encourage users to take more photos on the front end instead of having to comment on every individual observation.
My first tip would be, āPhotograph the sides of flowers,ā or some variation on that theme as that view can be very helpful on many of the taxa I work with. I know there is no way to come up with a complete list since every taxa can depend on different features for identifications, but I think even some broad tips might help. For plants, I usually take certain types of photos on any new taxon I encounter, and I imagine there might be some baseline set of images that might be helpful for other major taxa as well. The tips could even include ideas about how to get better photos with cell phone cameras as out of focus images still seem to be a big problem with autofocus devices. The iNat banner could rotate through a different short tip every week so that the ideas might get to users who are not even actively thinking about them.
I donāt think this is a very popular thing to say, but I do think it is the reality. iNaturalist has a very diversified user base, and some people just take a single photo and upload it. And they have very little interaction with iNaturalist beyond doing that.
Yes, there will always be users like that, but I also see many users who have submitted hundreds of observations that do the same thing. There can be many reasons for that, but perhaps some can be persuaded to take more photos if the message reaches them.
As much as your time allows, provide feedback to users in your area of expertise and some of them will improve their ābehavior.ā I try to leave useful comments, although sometimes all I say is āLeaves?ā and hope they get the message. Even so, I am in the top 100 of plant identifiers on this site and I still contribute poor observations (especially while hiking with my phone) Genus Eriogonum (wild buckwheats) from Bosque Nacional Los Padres, Atascadero, CA.
I do think itās worth exploring what mechanisms might really improve the quality of evidence included in a typical observation. Hereās a very speculative approachā¦
Imagine if iNat had a second AI-based assistant that was trained to help observers add the best evidence to their observations. Letās call the assistant Nat, because apparently all these things need to have names. Nat would likely use an entirely separate AI model from the current one. Instead of providing suggestions for what organism is in an image, Nat would be focused on making suggestions for creating a good observation. Iām thinking that Nat could be incorporated into both the web interface and the mobile apps. Some ways this might work:
You upload six photos of different organisms to an observation. Nat detects that thereās very little in common among the photos and asks if you want to split them into separate observations.
You upload five photos of a moth and one of a spider. Nat is smart enough to suggest splitting these into two observations, not six.
You upload a single closely cropped photograph of an Aster-family flower. Nat recognizes that organisms of this sort are much easier to correctly identify with additional photos and asks whether you have other photos you can include, such as a side view of the flower, a view of the whole inflorescence, a view of the leaves and an overall view of the plant. Iām thinking that this logic would combine some CV aspects (detecting that this is likely in a certain family and that the initial photo is of the flower) and some curated aspects (photos of these items are the most important to identify Asteraceae).
You upload three photos of a spider. Nat recognizes that this type of organism is difficult to ID to species without microscopy and suggests that you stick to a genus-level ID.
You upload a blurry image of a flower. Nat detects that you took it with a smartphone and itās out of focus. Nat provides some very concise advice on handling autofocus issues and offers a button to get an email with more detailed advice.
You upload several photos of a plant and the first one lacks metadata. Nat prompts you to enter a date and time and/or location or to copy them from another photo that does have these data. (No AI here, just logic.)
The timestamps or locations differ greatly between the uploaded images. Nat alerts you so you can validate these.
And so on.
TBH, Iām really quite averse to most AI assistants, and this would have to be implemented carefully to avoid annoying experienced users and interfering with usability. But I would actually be quite grateful for additional error-checking when I create observations (and especially when other people create them!)
Though this isnāt going to work for all users, I think itās a good idea, sort of a ātip of the dayā like you see on many apps and websites. Telling users that they will get better IDs if they take closeups from several angles would be an eye-opening motivator for some.
This question is making me rethink some of my reactions. When I see a comment asking the observer, āDo you have any more pictures?ā my gut reaction is, why do you even need to ask? Since when is the answer anything but no? But maybe I should take such questions instead as a hint to the observer that more pictures are needed.
The answer is āyesā surprisingly often, especially with newer users. I have encountered many cases where someone only uploaded one photo of an organism, even though they took 4-5, because it was the ābestā/most aesthetic one
Hmmā¦ well, thatās a bit different. Actually, I recently came across a variation of the question: āThe picture is a bit blurry. Do you have a clearer one?ā And Iām thinking, if they had a clearer one, why would they upload only the blurry one?
I agree that your example is a strange instance, but I also have @thebeachcomberās experience. Often people upload the best one or two photos, but any that are rather blurry they miss off. Sometimes when you say e.g. ādo you have any photos showing the legs betterā theyāll find one and say - āoh I didnāt upload that one because itās blurryā - but a poor photo from the right angle can make all the difference. Similarly, sometimes I can take 20-30 pictures chasing some critter around, but Iāll only upload a selection that I hope cover the right details. Not knowing what the right details are though means sometimes someone asks if Iāve got any others and Iāll add a few from the desired angle.
Leaving tips using text replacement
One thing that can be done is to make autotext comments with these educational tips, including a link to any tutorial on the subject.
The keyboard settings in iOS make this really easy, and I would assume the same is true for most other operating systems.
Hi, I am new here. I would love tips on taking photos! I have a bunch to download, tons of mushrooms ā¦some I have never seen before. And I have heard mushrooms are hard to photo. I had to put my Nikon 950 level with the ground. Instinct I guess, and close ups? I do have a set of lensesā¦???
I was waiting for the holidays to end because I had grandchildren here. Now I have some time to figure out how to download. I am NOT techy. Crossing my fingers! Hoping I dont mess up. Maybe in the morning when my brain is fresh!
I really like your #4. iNat encourages ID to the species level, but too many times it is impossible to do. And sometimes Family may be as good as can be done.
I share your frustration and have vented it in journal articles (e.g. https://www.inaturalist.org/journal/sedgequeen/81386-show-me-the-other-side ) which, of course, few people will ever read. Adding more time and complication to the introduction process would likely be counterproductive. Mostly, weāre just going to have to deal with this aspect of citizen science.
A weekly ātip of the dayā could be helpful and wouldnāt be intrusive for those who donāt want to read it. Good suggestion. And an easy one since iNaturalist staff have almost infinite free time. (not)
This discussion may inspire me to make a few ācopypastaā paragraphs about common problems, so I can easily add them to observations. Itās hard to word them so they come across as helpful (or even evidence of an idiosyncratic observer) rather than condescending.
Maybe I have not explored iNat enough, but is there an online ābookā on how to take good photos of plants, insects, etc?
I personally do not like to disturb or stress my photo subjects more than necessary, so, for subjects like insects, reptiles, etc. I donāt like to hold or prod them - I take photos catch as catch can. I donāt like to pick flowers or even leaves to get good photos. I also donāt pick mushrooms to get photos of the gills, etc. Just yesterday, on my wanderings through a field I was questioning if I was not doing more harm than good because animals may leave the area do to my occasional presence, although I donāt think my presence is as bad as those letting their dogs run off leash.
Having said that, it would be good to know what views are desired / essential to make IDs for various specimens.
Usually I can get good-enough photos of the underside of leaves, backs of flowers, etc., by simply twisting the relevant part. I can often get the underparts of mushrooms by putting the camera down on the ground and photographing slightly upwards.
That said, sometimes I do take the cap off the mushroom or detach a leaf or flower. But these things are often easy to avoid.
One thing I would to encourage users to do is to consider providing scale in photographs. This is not easy for birds and other animals, but works well for plants and slower moving invertebrates. The best way to do this is to include a ruler of tape measure in the shot. Other things which work almost as well are ballpoint pens, eyeglasses or, in a pinch, your own foot. Coins and banknotes also work well, keeping in mind that iNatters in other countries may not be familiar with your currency.
As far as it has been explained to me, this is one āleave no traceā precept we probably shouldnāt worry too much about. Mushrooms are fruiting bodies ādesignedā by evolution to spread fungal spores. If you pick a single mushroom so that you can turn it over and photograph the gills (and maybe the structure of the stipe), youāre not damaging the main fungal organism, which remains as a mycelium network in the substrate.
After youāve taken your photos, you might want to lay the mushroom gill-side-down on the ground nearby, so that the spores can still filter out over time. Iām not saying thereās no difference, but it seems likely to be quite minor in most cases.
In many cases, experienced mycologists are going to want to see photos of both surfaces, plus the stipe (and sometimes a spore print) to provide an accurate ID.
Search in the forum for tips on what to photograph. (Iād do it for you and provide links but the coffee hasnāt kicked in yet this morning!)
Thereās no one place on iNat where all the tips on what to photograph are gathered. I bookmark sites where I do find tips on what to photograph and how to ID certain species, but I think the best thing to do is to concentrate on a particular group of species and learn (probably from field guides) what you need to photograph in order to ID them.
Hereās a general example of that: Many lichens specialize in growing on a particular type of substrate. Some grow on tree trunks, some grow on rock, some grow on the ground. While I am definitely not a lichen expert, I do try to take a close-up of a lichen plus a larger view that shows the whole lichen on its substrate. Even those two photos arenāt usually enough to ID a lichen to species, but it will often get the ID to genus.
A more specific example: American and Beaked Hazelnut, two common shrubs where I live, are very easy to tell apart when they are fruiting, but they donāt fruit every year and smaller, more shaded shrubs may not fruit at all. So if a hazelnut is fruiting, I take a photo of the fruit. If one isnāt fruiting, Iāve learned that you can tell them apart by the presence or absence of tiny hairs on the twigs, so I take a close-up photo of the twigs. Learning little details like this takes time and I often forget from one year to the next what I need to photograph for uncommon species.