Depending on what you really need, you might have different options. I don't think collar/bracelet for great apes is currently resolved reliably, and could pose risks to the animals. Implantable transmitters are one option but more invasive. I think Chris Walzer (currently at WCS) has done some initial work on implanting orangutans (or gibbons), but that was quite some years ago. Others might have done it since as well. FWIW we have implanted captive chimpanzees with subcutaneous transmitters for other reasons and some have removed/damaged it, so that's a risk. Intraabdominal might be another (though even more invasive) option. 

Is  visual observation an option? This is well used in many species, though might be very costly/labour intensive. I have been involved in a project with howler and spider monkey where this has been used successfully, happy to put you in touch with them. 

If your questions is primarily absence/presence of certain individuals, you could also try DNA based methods? And someone suggested below individual ID from video/pictures, this has been done in captive great apes, so it is an option, but you still need to get the shots in the first instance, so probably not many advantages over manual monitoring methods. 

Our city-neighbours Robotto are a Drone AI-software company and have an ongoing animal tracking projects in Thailand, Australia, and Greenland (probably more by now) in co-op with WWF, using drones. 

Give them a look! I know Kenneth, their CEO, pretty well so can match you two.

A rough explanation of how the tracking process happens is:
A forest ranger brings a suitcase with a drone to a watchpoint, pilots the drone around the area for 30-50 minutes while monitoring detections real-time on the provided screen.

Have seen it live in Thailand, it was impressive! :)

In a recent call with researchers in Thailand. They have mentioned that they use passive chip readers to log data about chipped animals as they pass by.  

Great resources being shared! Darwin Core is a commonly used bio-data standard as well.  

For bioacoustic data, there are some metadata standards (GUANO is used by pretty much all the terrestrial ARU manufacturers). Some use Tethys as well.

Recordings are typically recorded as .WAV files but many store them as .flac (a type of lossless compression) to save on space. 

For ethics, usually acoustic data platforms with a public-facing component (e.g., Arbimon, WildTrax, etc.) will mask presence/absence geographical data for species listed on the IUCN RedList, CITES, etc. so that you're not giving away geographical information on where a species is to someone who would use it to go hunt them for example. 

Hello, I am experienced in conservation data strategy. If you want to have a conversation you can reach me at SustainNorth@gmail.com.

Bird Monitoring Data Exchange is a standard often used for birds data. 

 Is thisvfunding grant an opportunity? https://www.dronedeploy.com/blog/expand-your-impact-with-a-grant-from-dronedeploy

Hi David, this is an incredible project. Would you be interested in sharing more of your experience with AI and wildlife conservation with my students? They are currently researching this, and would greatly benefit from speaking with a professional in the field. Thank you for considering!

Hats off to your team for this absolute game-changing technology! 

We rescue stray and wild animals in Taiwan, and the bulk of our work is saving animals maimed by wire snares and gin traps. We've become better at finding the devices, but still not good at all. There's simply too much difficult terrain to cover and we only have eyeballs and hiking sticks to find them. We know roughly where they are because the maimed stray dogs will eventually find their way onto a road and be reported to us. Then we close one of them, set up a trail camera, get the evidence of the poacher in the act of re-setting it, and get him prosecuted and shut down. But we need to be able to scale this greatly.

I've been using a thermal-imaging drone to locate stricken animals and am now considering buying another drone more suited to finding traps and snares. Some newer drones are able to navigate through forest without crashing into thin branches, so I've been looking into equipping one with LiDAR to see if that can detect the devices. But then I came across your YouTube channel and then this post about using airborne synthetic radar, and I'm incredibly excited to see where you might take this incredible technology.

How can we get our hands on the SAR you're using? It's 3 kg, right? I'm wondering if I could fit it to a suitable drone. If it works above forest canopy to detect traps and snares on the forest floor, then I can use a load-carrying drone instead of a light obstacle-avoidance drone.

If you made the SAR yourselves, then maybe think about crowdfunding for your project. I'd happily pledge funds if it meant I could get my hands on the kind of equipment you're using.

I can't tell you how happy I am thinking about all the animals' lives you'll save with this. Don't just remove the snares—gather evidence and put the poachers out of business too!

Sam  any update on Instant Detect 2.0 - previously you mentioned that you hope to go into volume production by mid-2024?

I would love to also see a comparison between Instant Detect 2.0 and Conservationxlabs' Sentinel products if anyone has done comparisons.

Are there any other similar solutions currently on the market - specifically with the images over LoRa capability, and camera to satellite solution?

As others have said, pretty much all image models at least start with general-subject datasets ("car," "bird," "person", etc.) and have to be refined to work with more precision ("deer," "antelope"). A necessity for such refinement is a decent amount of labeled data ("a decent amount" has become smaller every year, but still has to cover the range of angles, lighting, and obstructions, etc.). The particular architecture (yolo, imagenet, etc.) has some effect on accuracy, amount of training data, etc., but is not crucial to choose early; you'll develop a pipeline that allows you to retrain with different parameters and (if you are careful) different architectures.

You can browse many available datasets and models at huggingface.co 

You mention edge, so it's worth mentioning that different architectures have different memory requirements and then, within the architecture, there will generally be a series of models ranging from lower to higher memory requirements. You might want to use a larger model during initial development (since you will see success faster) but don't do it too long or you might misestimate performance on an edge-appropriate model.

In terms of edge devices, there is a very real capacity-to-cost trade-off. Arduinos are very inexpensive, but are very underpowered even relative to, e.g., Raspberry Pis. The next step are small dedicated coprocessors such as the Coral line (). Then you have the Jetson line from NVidia, which are extremely capable, but are priced more towards manufacturing/industrial use. 

Read "Products | Coral" on Coral

Products | Coral

Helping you bring local AI to applications from prototype to production

Coral

I am certainly a somewhat vertical thing with a lump on top, so I get it...