• BirdForum is the net's largest birding community dedicated to wild birds and birding, and is absolutely FREE!

    Register for an account to take part in lively discussions in the forum, post your pictures in the gallery and more.

Will image recognition revolutionize birding? (1 Viewer)

jurek

Well-known member
I had an interesting thought.

Image recognition is used to everything from identifying faces of friends to cancer nodules. What if somebody develops a software which recognize and identifies birds on photos, and then it is paired with automatic CCTV cameras?

Imagine CCTV cameras set in reserves in remote beaches, pools or reedbeds, where people don't often visit. Or maybe automatic feeder camera recording birds when the owner is away from home. Or cameras in remote rainforests, islands or desert pools. Whenever something triggers a motion sensor, several photos will be taken, software would identify the bird and send a photo with a message via a mobile-like link.

Or imagine automatic camera stations left in the field for a year, with software automatically identifying birds and presenting a set of species for confirmation.

Software will not be foolproof. However, it is enough when a software recognizes at least there is a bird and sorts species with say 90% accuracy. This would sort pictures of bird species for a human to quickly accept or reject. This could do work which ornithologists could not process otherwise.

Imagine how much information about rare birds would be gathered. And how many rarities would be found and later twitched in places not well watched!

And if ornithologists might be afraid thier skills would become obsolete? Don't worry. Similar projects in other topics show that data grow so much, that although mundane tasks are automated, specialists have more work than ever.
 

jurek

Well-known member
One can discover several times more vagrants even in a well covered coastal spot, if there were automatic camera inside places which are not accessible, e.g. inside mudbanks or reedbeds, in bushes growing on steep slopes, in bushes in the middle of pastures etc. Every bird reserve has such inaccessible places.

I am wondering how this concept can become real. It can revolutionize bird surveys in the tropics. But it might first become developed and profitable for mundane feeder and hunting cameras.
 

temmie

Well-known member
Obsidentify (app related to observado), but it’s beta version and I don’t know hoe good it is. Also for plants and insects. I thought inaturalist has something similar...lots of work going on at universities at the moment.
 

Hauksen

Forum member
Hi Jurek,

Or imagine automatic camera stations left in the field for a year, with software automatically identifying birds and presenting a set of species for confirmation.

Coincedentally, I just came across automatic bird recording in a technology-related forum.

It's quite impressive what the user managed to learn from a bit of cheap Arduino-based techology about the behaviour of the Swifts nesting on his house:

https://www.arduinoforum.de/arduino-Thread-Auf-SD-Karte-registrierende-Waage?pid=3294#pid3294

Direct link to the graph (not sure that this will work):

https://www.arduinoforum.de/attachment.php?aid=322

It's in German, but if you click on the small thumbnail in the middle of the post, you'll see a graph showing last arrival and first departure time at the nest per day, along with sunset and sunrise, and number of feeding flights per day (which topped out at 25 when the young were freshly hatched).

That was based not on image recognition, but on recording scales at the nest entry (so there's actually even more data available than shown in the graph), but I think it's a great example for just "dense" automatically recorded data can inform us about what birds are doing.

To test the scales, the user over there put them next to a feeder, and so he could record the weight of each of the tits landing there. Combined with image recognition, you might be able to monitor the health status of individual birds pretty much in real time ... and that's just scratching the surface.

Regards,

Henning
 

Andrew Clarke

Well-known member
Look forward to the days when AI can accurately identify large white-headed gulls and their hybrids. Would save us mere humans a lot of head scratching!
 

PYRTLE

Old Berkshire Boy
But surely the use of the word in the post heading " birding" means to actively search for and identify a bird by its call or plumage, through knowledge and experience. This to me includes research and referring to field guides, bird recordings and images. Otherwise the correct word for just audio proof or visual proof then becomes "recording" a species. Almost like a speed camera,.....trail cameras have previously been discussed.
P
 
Last edited:

jurek

Well-known member
Well, once we know a rarity is there, then go and see it. I imagine that a rare bird, say a warbler, is first observed on a remote camera in some hidden corner of a bird reserve. Re-finding and seeing it in person would still be a challenge.

In a sense, remote cameras might work as an overchanged rare bird alert.
 
Last edited:

PYRTLE

Old Berkshire Boy
Well, once we know a rarity is there, then go and see it. I imagine that a rare bird, say a warbler, is first observed on a remote camera in some hidden corner of a bird reserve. Re-finding and seeing it in person would still be a challenge.

In a sense, remote cameras might work as an overchanged rare bird alert.

In my active birding time ( starting about 1966, age 11 years), birding for me had been finding and identifying birds with eyesight, ears and binoculars. Later on , in my twenties I bought a telescope and tripod so I could see birds at a greater distance and magnify the image. This is how birding was for me and many others of my age group for decades, backed up with a field guide and a notebook. Some took photos and many " firsts" were recorded with black and white or colour wet film photographs. The serious guys had bulky tape recorders in the field. Hand written field notes and sketches were the norm and for a long time the way to document sightings, even for county and BB records.
This is/was the birding world for me and many others from the 1970s onwards ( till the advent of Nancy's cafe and Birdline ).
Nowadays, there is the digital camera world enabling superb images for scrutiny and records.
I look forward to when a first for the UK is found by a trail camera or A.I. / recognition binoculars.
P
 

Farnboro John

Well-known member
When I'm not birding (and sometimes when I am) I spend time watching aeroplanes. In my time I was a number-cruncher, a reggie-spotter. Not any more, I take aircraft photos to try to record the essence of their flight, shape, the play of light on their surfaces.

Meanwhile Mr. Reg S. Potter has moved on.... even the old boys brought up on notebooks and brass-bound telescopes sit at the end of the runway at Farnborough with their tablet running FR24 and suddenly erupt from their BMWs to look up and see G-BNLY at 30,000 feet as a tiny speck at the head of a blossoming contrail. Dot-spotting! Never miss a serial - but never read one off a speeding jet as it whips past at 450 knots....

What's missing? Challenge.... why do we do hobbies? Is it really all about the numbers? The tick? Or is it the journey to competence, working it all out in the field individually and also, importantly, having to get the information in the first place. Up until we stop having to use fieldcraft to obtain position, observational skills to see features, camera skills to get the picture with the feature that moves too fast (but still observe the image and interpret it). Until we stop making it our own mission and give it to a fixed sensor or a drone.

Perhaps these things are a great way to get data to advance conservation and science, but for me, so far as birding is concerned, to hell with them.

John
 

Sancho

Registered User
Supporter
I'd rather take my eyeballs out with a spade than let image-recognition replace my binos and (what's left of) my brain!;)
 

AntonBE

Well-known member
I think a lot of people here are seeing image recognition as a replacement for our birding efforts, instead of being a useful tool that can be used (or not used) by anybody. Birding as a hobby will always stay that way, the people seeing the birds (us) will always still want to see the birds with our own eyes or hear the birds with our own ears.

Yes, identification is currently a significant part of the fun in birding, but I don't see any of these tools ever completely getting rid of in situ identification altogether. Not every bird will let you photograph or record it, and not every bird will be close enough to photograph or record. Photographing and using an app to identify every single bird you see will never be a viable and efficient birding strategy; Even if systems like image recognition make identification to a species level easier, you would still have to recognize something is special and worth photographing in the first place.

I don't really see how image recognition apps and websites can ever be 'bad' for birding. Take bird guides, for example. Those have been improving with more and more detailed descriptions and illustrations over an ever growing range of countries and continents: I don't see anybody complaining about bird guides making birding less of a 'challenge,' even though that is definitely what they do. A website or app that can notice details from a picture and help you with identification is just another "tool" that us birders have the option to use - and that's a positive evolution in my book!

The way I see it is that systems like these have the option to make birding a much wider hobby that reaches out to more people... And that's always a good thing!

Trail cameras paired with image recognition, as suggested in the original post, would definitely be interesting.
 
Last edited:

Farnboro John

Well-known member
Access to good and improving information on how to identify birds is a world away from something doing it for you.

When I was young, everyone knew their friends' phone numbers. Now nobody does, because those numbers are at one remove: nobody goes into their phone and looks for a number, they scroll down and press "Eric" or "Erica". Not only does that mean when their phone runs out they can't dial on anyone else's, it means they aren't stretching their own brains. The road to drooling idiocy....

Field Guides, yes, apps, yes. AI, no. Not even for LWHG!

John
 

etudiant

Registered User
Supporter
The advance of image recognition is unstoppable, the hardware and the software both are getting huge investments for other purposes, with birding a minor application, useful for product development.
So the task for birders is to avoid acting as luddites, but rather to channel this emerging capability for useful purposes.
Jurek made some helpful initial suggestions, which seem of lasting worth imho.

We go birding because the birds are delightful to watch. The lists are really a diversion, they even distract from the joy of observing and documenting nature when they become an end in themselves rather than a tool for monitoring abundance and cyclical variation.

Anyone who manages a patch list will immediately recognize when something peculiar shows up. It does not detract if the ID is made by an app rather than by the observer. There may be instances of people losing interest because they miss the thrill of discovery, hopefully such cases will be rare compared to the number of people who are awakened to the diversity around them.
 

Farnboro John

Well-known member
The advance of image recognition is unstoppable, the hardware and the software both are getting huge investments for other purposes, with birding a minor application, useful for product development.
So the task for birders is to avoid acting as luddites, but rather to channel this emerging capability for useful purposes.
Jurek made some helpful initial suggestions, which seem of lasting worth imho.

We go birding because the birds are delightful to watch. The lists are really a diversion, they even distract from the joy of observing and documenting nature when they become an end in themselves rather than a tool for monitoring abundance and cyclical variation.

Anyone who manages a patch list will immediately recognize when something peculiar shows up. It does not detract if the ID is made by an app rather than by the observer. There may be instances of people losing interest because they miss the thrill of discovery, hopefully such cases will be rare compared to the number of people who are awakened to the diversity around them.

On the bright side, its the end of the self-found nonsense! :t:

John
 

Andrew Clarke

Well-known member
The advance of image recognition is unstoppable, the hardware and the software both are getting huge investments for other purposes, with birding a minor application, useful for product development.
So the task for birders is to avoid acting as luddites, but rather to channel this emerging capability for useful purposes.

Well said Etudiant. Interesting how technology always influences ornithology. The developments in optics (from binoculars to telescopes), photography (film then digital cameras/video), sound recording and playback (from cassette tapes to field recorders and automatic digital listening stations), identification guides (from books to apps) and now the developing AI identification software are just a natural evolution and will no doubt continue to lead the way at the cutting edge of this hobby and science.

I’m old enough to have lived through and witnessed the astonishing rise of the internet and find it hard to get my head around and truly appreciate the incredible changes that have occurred in such a short time. It would be fascinating for some of the younger generation of birders to have a go at birdwatching 70s or 80’s style!

One thing’s for sure though; much as I enjoy the fruits of the digital revolution, that basic raw connection with nature which I experience when out and about (sometimes even without optics!) is more profoundly important than ever.

Good Birding

Andrew
 
Last edited:

Sancho

Registered User
Supporter
I have zero interest in looking at any bird through a machine that identifies it for me. I want to 'see' birds with my own eyes, maybe through some glass bits, and do all the thoughts, processing etc. with my own brain. Same as reading a book in a language I can speak, or read, even with difficulty, rather than in translation into my first language. As for 'AI-Birding is almost upon us', well, IS-birding in a useable format for human hands has 'almost' been upon us, hallelujah, for two decades. My breath remains unheld.

P.S. Kinda !ike Viagra- you don't 'need' it, you just need to re-pioritise your interests;)
 
Last edited:

Users who are viewing this thread

Top