I'm looking forward to the cool Margo TV we are going to see when we get some AI bear monitoring going in Canada with Margo supplies :) That's going to be awesome.

Here is an image of a bear from last night in the Scandinavian zoo in Denmark. The system triggered on the bear even though it is very dark and this is a digital zoom on a wide angle fixed camera. Despite being very fuzzy we are able to get triggers on these ones due to the use of multiple mitigation techniques so we get long distance, high sensitivity and low false positives. This one for example was checked with multiple successive frames and multiple AI models before calling it a bear.

This is without thermal. But with added thermal and models trained on thermal images the heat signature will effectively be taken up in the model I'm expecting. This would mean that in effect you can tolerate a lower confidence level but it would after all be a living thing as well.



This trigger event for this event was even less clear, but you can still just make out two eyes, and bear in mind (excuse the pun), the actual view from this camera is even wider and the bear further away in a sense as this is a dynamic zoom.

You can hack a 6-inch underwater speaker onto a recording unit, and swap it from being  speaker to a microphone. They are, on paper, the same thing.  That frequency range doesn't work will with small piezo surface (itty microphones). If you want to get fancy, you can add the resistors and capacitors to make a band-pass filter, so you don't pick up quite as much thunder and passing vehicles.

This is a fairly standard audio problem that a good audio engineer can assist with.

I've detected low frequency bird wingbeats with some low-power analog wizardry in the past. But you will need dedicated audio engineering time.

Hi Folks, 

I'm Liz Ferguson from Ocean Science Analytics and just noticing this thread so reaching out to express interest! Thanks for posting that overview Stephanie. 

My company is an ocean science research consulting and online technical training company. We provide hands-on training for software, thus far related to passive acoustic monitoring and R, but have a Marine spatial analysis course coming out using QGIS in 2023. I also teach at a local college, and some analytical courses. Really interested in the group so happy you started it Suz and Andrew!

Hi! I'm Connie, I'm in biomedical and have spent the past few years in Texas, working between Academia and private sector/NGO. I've been coordinating open source technology development between multiple institutions, and private sector, where I create technology and apply it in the field under for profit and non-profit labels. Generally my day is up at 6am, do tech until lunch, see friends and family for lunch, do applied bird work and training, have dinner, do tech, see friends and family in the PM, with field work trips in between.

My part-time project work under USDA Conservation Innovation Grants ends in March, freeing me up 20 hours a week to pursue pure tech development and publishing, and I'll actually have time to say hi to people again!

My biggest interest is bringing in and creating interdisciplinary early career conservationists and training self-made conservationists who have created their careers in areas with few resources and support. Sometimes, one just needs a few critical skills to get their projects off the ground. I mentor research students as well as applied conservation learning using 2,000 captive birds at a propagation aviary outside of Austin, TX, where I'm the Institutional Officer overseeing research for the IACUC.

If you need to practice collaring some birds before you do it in the middle of nowhere and spend the travel money, let me know! My bird research phone number is US and easy to remember 42 42 MAGPIE. (That's the meaning of life twice, and a magpie.)

I've hosted trainees from the Field Museum, Natura Mexicana, a load of universities in US, Europe, and South America... you do birds, you do tech, let's be buddies!!!

Hi! Did you find all the collaborators you needed?

I'd be curious to hear what sort of fruits were harvested from your collective bioacoustic efforts!

Scott in Seattle

It sounds like you've exhausted everything exiftool can do, in which case,  YMMV, but I have a module that uses a combination of OpenCV and PyTesseract to pull metadata from the actual pixels:

https://github.com/agentmorris/MegaDetector/blob/main/megadetector/data_management/ocr_tools.py

Documentation for this module is here:

https://megadetector.readthedocs.io/en/latest/data_management.html#module-megadetector.data_management.ocr_tools

I have only really used this to extract time and date; my goal was to make it as robust as possible for time and date over a wide variety of manufacturers, rather than to find a wide variety of text.  But I expect it would be relatively straightforward to tune for a single bit of metadata and a single manufacturer.  I'm not sure whether it's worth literally using this module vs. just looking at the code and using that as a starting point,  or using the similar R example you linked to, but FWIW, if you try it out and get stuck, send me a few images by email and I'll see if I can tune it a bit to get what you need.

Good luck!

-Dan

Hi Adrien, should the letters of support be written by team members or by external supporters (researchers from other institutions, local organization managers, etc.)?

Ok so I'll add my 2-cents here but only really in relation to the 'corrosion in seawater' part of the discussion. 

When I am releasing packages, sensors or moorings sub-surface I'm using (as has been previously mentioned) more expensive burned wire units that can be programmed for particular durations and then burn a wire to release or you come along and actively talk to the burn wire unit with a hydrophone/transmitter unit and sent it a burn signal when you are close by. 

One other option I found is to have galvanic links made up. These are small cylinders of alloy that are made up to corrode and degrade over time using a special recipe and a calculation based on your depth, water temp and salinity. You can get a bag of 50 of these links for not a high cost but they may only last for 2 weeks and the corrosion/release accuracy could be in the region of 4-12hrs I believe. Still its perhaps an option for things that are more static and not moving around into differing temp/depth zones.

The galvanic links I have been using are found at:

Read "Galvanic Timed Releases - International Fishing Devices" on

Galvanic Timed Releases - International Fishing Devices

For over 40 years, International Fishing Devices, Inc. has been the designer and manufacturer of Galvanic Timed Releases.

Some thoughts as I have experience working with some of the tech mentioned...

Corrodible pin

@htarold Did a great job explaining how that works. This pin is used in the pop-up satellite tags from Wildlife Computers, Microwave Telemetry, and Lotek Wireless. Biologging Solutions in Japan is also making a PSAT (I assume they also use a corrodible pin) and Star Oddi is also trying to develop the same. In short, nearly everyone uses the corrodible pin. 

Burn wire release

There is a patented design (Desert Star Systems) that uses burn wire (I forget the type of wire), and if my memory is correct, it can work in any environment (perhaps I'm wrong, but I'm pretty sure it works in air, too). You can Google that company to get more information, and they used to have some YouTube videos showing their burn wire design. 

Propellant-based release

I wanted to say 'explosive' release, but that's being a little dramatic. There is a pop-up satellite tag manufactured by Desert Star Systems that uses an explosive-based system (there, explosive, I said it). Instead of a pin, their nosecone has a chamber that you load gunpowder into, and upon release the gunpowder is ignited, causes combustion, the chamber cannot support the expansion, and the nosecone is released from the tag through force. It works in any possible environment. 

Galvanic timed release (GTR)

In my opinion, the biggest problem with GTR is that the timing is variable. It is heavily reliant on the water quality, and I have heard anecdotally that it is near impossible to truly know when the device will release. 

Unknown release

I know that Wildlife Computers was awarded some funds from the Office of Naval Research to develop a release package for recovering instrumentation from marine mammals when they haul out (think elephant seal). I believe this package is independent of a tag so you can use it as a carrier for other instrumentation. I don't know whatever came of this project, but it's Google-able. (Update: it's called the PRD-RP, and if you Google that you'll find it easily enough.)

Disclaimer: I used to work at Desert Star Systems, and consequently I know their releases well. Nothing I shared here is proprietary or a trade secrets. 

Hi, likely already a bit late to make impact, but here's what we have seen work very well:

• UAV's equipped with thermal vision
• Fence alarms
• 'Scarecrows' mimicking ranger activities with a pivoting flashlight (very effective and cheap)

We are a NGO developing those cheap, but efficient methods and partnering with a game reserve. Currently looking into scale up to make the tech available to other reserves.
I haven't been involved as a founder, but a partnership with the reserve was always required.
The goals a shortly put as, preventing wildlife crime BEFORE it happens. So mostly any kind of poaching. Now as we are growing, looking also into monitoring tasks such as wildlife count/VHF tracking etc.

You can also check out WWF's newly launched C2C: Conflict to Coexistence approach. It provides a framework and methodology for developing tailor-made human-wildlife conflict (HWC) management strategy based on the local cultural, environmental and social context.

Read "" on