Grupo

Computación de borde / Canal de noticias

Este grupo reúne a quienes trabajan en la intersección de la IA de borde y la conservación, centrándose en el procesamiento de datos en tiempo real y en los dispositivos para la monitorización ambiental, con el fin de facilitar el intercambio de herramientas, modelos y estrategias para superar los desafíos en áreas remotas y de baja conectividad.

discusión

Mini AI Wildlife Monitor

Hi All!I've been working on various version of small AI edge compute devices that run object detection and Identification models for ecological monitoring!I've recently been...

27 14

In case someone. Find this totally out of place commemt… this is how I solved it, I've decided to use a IMX477 HQ Camera, building a *manual, heavy-duty optical rig* utilizing C/CS-mount lenses and physical macro extension tubes.

Wow, what a great project.

This is a great project! Some comments:
RaspberryPI though accessible is not the best fit for video pipelines and AI workloads or off grid deployments:
- it lacks onboard ISP which means either software implemented ISP, distorted data or on camera non energy optimized ISP.
- it lacks any power management techniques, low power modes, etc.
- it runs from SDCard using the same one for OS, swap and data, any corruption can lead to full loss.
- it runs any AI/ML workload on CPU which is extremely non efficient and any addon accelerators such as Hailo8 add a lot to power consumption and heat dissipation representing more challenges.

The advantages are of course plenty of documentation, community and all kind of makers addons, hats, etc.

For something more realistic, real life suitable I would suggest using something based on SoC with integrated NPU such as Hailo 15, Renesas RZ/V, Synaptics SL1680, MediaTek Genio or even the I.MX8M Plus for very light AI/ML workload. All of these have variety of SBCs, kits or even standalone smart camera oriented designs available from different vendors.

Ver la publicación completa
discusión

Raspberry Pi Alternatives for Edge AI!

Hi all!I've been recently looking into Non-Raspberry Pi (and Jetson) SBCs that have some form of Neural Network acceleration on board. I'm trying to find a good balance between...

1 2

Hi Luke,

Really enjoyed another one of your videos, you've definitely got me experimenting more with edge compute and AI in my garden. I was wondering whether you’ve done much with NVIDIA Jetson boards? I think you mentioned having an older version in a previous video.

I got an Orin nano super developer kit since the higher-memory RPis doubling in the past year, while the Orin Nano/Orin Nano Super now seem surprisingly fairly static in price over that time period. What interests me is the possibility of using them for wildlife monitoring: running detection or ID models locally, collecting new field data, self labelling and training then using that data to periodically improve the model all on the edge.

I’m not sure whether full on-device training would be practical, but perhaps a more realistic workflow would be active learning or a teacher–student/pseudo-labelling setup: the device flags uncertain or interesting detections, those get reviewed or labelled, and the model is retrained or fine-tuned periodically self redeploying via scripts or a local model and agentic harness.

There also seems to be a good ecosystem of cameras and sensors around Jetson. I assume you find limitations in the Jetson boards for ecology/wildlife-monitoring projects?

Ver la publicación completa
discusión

AI Edge Compute Based Wildlife Detection

Hi all!I've just come across this site and these forums and it's exactly what i've been looking for!I'm based in Melbourne Australia and since finishing my PhD in ML I've been...

25 4

I find the performance of micro-nano sized models that run on MCUs impractical for many applications. This is due to the low FPS, tiny Image resolution processed and very low model capacity.
I think people underestimate the huge jump in complexity from something working on the benchtop detecting a person from a meter away to trying to detect a cat-size object several meters away in a noisy environment.

Ver la publicación completa
discusión

remote sites as 'Edge computing'

First and foremost, a HUGE THANK YOU!! to the Wildlabs team that put on a fabulous ICTC 2026 event in Lima ... the energy and diversity was amazing! Secondly, a HUGE thank...

17 1

Oh yeah. On a Jetson that sort of thing would be viable.

Our Pi 5 is really pushing the boundaries with what we have it doing.


To put it into perspective. Our polar bear detector is round robin checking 22 cameras with a more than 20,000,000 parameter model on a middle of the range Jetson.

Yes, I was considering inference only in this case. I see it as a filter applied to raw images and audio before uploading to cloud/server to save bandwidth (if it is necessary at all?), which I think was the OP's request. 

It is then a balance of power requirements and power stability (and physical security?) to keep a small server running and connected. But a starlink system is relatively power hungry for a remote station too, so they may have this covered.

Software wise, the "TRAPPER" system might do what you need, with a local/LAN installation and allow for admin access externally. https://os-conservation.org/trapper/

 

nothing like a concrete, interesting problem to find focus
:-]

my partner is into chickens ... trust me, it's a whole subculture 
🤣

We were talking with friends about pecking orders , and thought... hmmmm

so, after riffing with Claude a bit, assembling RPi4+ USB Coral acclerator + Tapo 120 to tinker with RTSP streams Frigate Plus , COCO, Pytorch, etc. to automatically determine a pecking order, from an RTSP webcam.
Claude's got the strategy all figured out! 🤣

And we've got an Audiomoth and a StealthCam we use on occasion.

SO, managing/evolving that RPi4-based from behind the firewall at her house, versus mine, will be fun. 

Just installed the Wireguard client; off we go...

(Viv, top of the pecking order, sole survivor of a previous weasel attack; all the others are newcomers...)


 

Ver la publicación completa
Link

Solar Powered Mini AI wildlife Monitor!

Hi all, I'm sharing the latest update to my edge-compute wildlife monitor, this time Solar powered for full off-grid operation! The Code and 3d printable STLs files for this project are all available!

4
carreras

Ecological Data Scientist

The Smithsonian Institution is the world’s largest museum, education, and research complex, with 21 museums and the National Zoo. This position is located in the Smithsonian's National Zoo and Conservation Biology...

1
Ver la publicación completa
discusión

Help us find our cover image!

Hey everyone! We're giving our Wildlabs group a proper visual home and we need your help!We're looking for a cover image that truly captures the spirit of edge computing in...

6 0

How about this one. Three wolves captured on a solar powered thermal imaging triggered wolf detector. One wolf about to jump the ditch to take a closer look at the horses in the field.

This marks the end of an era. The alpha she-wolf of this pack was unfortunately killed on a highway today. She raised many pups. We grieve for you Noella.Three wolves from Noella's pack
The video for this shot is here

 

Hi Philip,

Here's an overview and a blog on Voltaic's website. We manually recover the microSD cards, but it would also be possible to use Iridium SBD by including a Rockblock connected to the RPi for full recovery of images (although we don't need to, as the ideal situation is to remove the cameras and leave nothing behind many years later). 

Ver la publicación completa
discusión

Seeking input: FAIR & AI-ready wildlife drone datasets

Have you ever tried to reuse a published drone dataset, only to realize key context was missing? Or wanted to publish your drone dataset, but not sure how to start?We are seeking...

4 5

Thanks Sean! Looking forward to the community's feedback on this! Feel free to fill out the survey, or open an issue on GitHub to track specific suggestions.

Hi @jennamkline , are you in the SCB Drones and Data Working Group? I think it would be great to have that as a topic for discussion during the next meeting! I'd be keen to hear more about this work.

@annavallery this might be of interest to you.

hi @elsa thanks so much for the suggestion! i would love the SCD Drones and Data Working Group's input on this proposal. I was planning on attending one of their meetings later in March to solicit feedback. If there are additional venues to engage with the group, please let me know!

Ver la publicación completa
evento

Edge AI for Conservation Workshop

Deploying Intelligence Where It Matters Most

2 8

Hi All, thanks so much to everyone who participated in the workshop and made it a success! We've uploaded the slides here: 

Thanks!

Jenna 

Hi All, thanks so much to everyone who participated in the workshop and made it a success! We've uploaded the slides here: 

Thanks!

Jenna 

Ver la publicación completa
evento

WILDLABS Edge Computing Group Meetup

Connect with conservation technologists innovating at the edge — and help shape this new WILDLABS community.

3 1

Hi @jennamkline and others,

There's some confusion on the time for this event. It's listed as November 4 at 11:00 am ~ 12:00 pm EDT (UTC -4), but daylight savings ends in the US on November 2. November 4 is during standard time. Easter Standard Time is UTC -5.

Is this meant to be 11am to noon EST?

I see the event image says 11 AM EST. I'm on the West Coast, and the Zoom page said 8 AM Pacific Time (no mention of standard vs daylight savings). After I registered, the calendar invite is showing up as 7 AM.

I believe the call will be at 8 AM PST, so that's how I have it on my calendar, but I suggest people double check!

Ver la publicación completa
Link

Solar Powering a Raspberry Pi!

I've been building a lot of experimental devices for edge computing. After not being able to to find a great solution for powering a Raspberry Pi, or other SBC, from a solar battery set-up that could also interface and communicate with the Pi we decided to create our own!

1
discusión

Connecting with Conservation Tech Leaders in Kenya

Hi everyone! I’m currently in Nairobi, Kenya for a few days and would love to connect with fellow group leaders or community members working on conservation technology. It would...

3 0

Hi Youssef

Greetings from Arusha, Tanzania. We are happy to see that you are interested in connecting. I am a co-leader for the East Africa Group. I would like to connect and invite you for a tech-talk with our fellows, East Africans. If you have got some space, let me know.

Hi Susan , 
Thank you for your reply and your kind invite . Unfortunately my stay in Nairobi was short but there is a chance to come back to visit east Africa again in the beginning of November and potentially going to Tanzania . Let's keep in touch :) 

Ver la publicación completa
discusión

Solar edge-AI forest watcher (prototype)

I’ve built a small, solar node that listens for chainsaws/vehicles/gunshots, snaps one photo on a confirmed event, logs to microSD, and sends a tiny alert over Wi-Fi. It runs...

1 3

Hello Samuel,

First of all, all the best of success with the Edge AI Earth Guardians contest.

Based on my experience with using PAM for poaching/hunting pressure monitoring, a photo would not be really useful - unless you have a very specific setting (which I can't think of right now). The strength of PAM is that you do not need to have line of sight to detect a sound event (unlike with camera traps). Now, if you were detecting vehicle sounds, and you had the camera at a road bent, then yes, in principle, a sound + image could be useful (but then again...the sound will trigger the sensor long before the car/truck appears in the field of view of the camera....).

My concern with real-time AI is how to eliminate false positive triggering (esp. with gunshots sounds). Unless this is resolved, I would be skeptical about the use of similar solutions (and we have tried something similar in the field). 

Finally, another concern is the way of data transmission - here you are using WiFi. So this is to be used at a homestead - farm setting, and not really at a remote forest, which is where we most need early warning systems.

Best regards, Christos 

Ver la publicación completa
discusión

Exploring the Wild Edge: A Proposal for a New WILDLABS Group

Over the past year, I’ve found myself returning again and again to one big question: How can we make conservation tech work where the wild really begins , where the signal...

60 27

Hey everyone!
This looks really cool and I'm really interested :)

Remote sensing is indeed one of the major challenges in conservation, especially with the rate and diversity of incoming data. I recently attended a climate codefest where the goal was to compress this incoming data using AI to make sense of it. We ended up using unsupervised learning to cluster meaningful data and transport it on limited bandwidths.

But there is so much more to do and learn! I'm glad you guys made this group. I would love to explore more of this area and learn from your experiences. 

How can I be more involved? Let me know!

Cheers, nikhil

Hey Nikhil,

Thanks a lot for your message! The group is no longer just a proposal , it’s now officially created 🎉 You can join us here: 

 

I’d be super keen to hear more about your codefest experience , it sounds exactly the kind of knowledge-sharing that could spark great discussions. Why not share it in the group? You’re more than welcome!

Looking forward to learning from you, :)

Youssef 

Ver la publicación completa