@holmbergius Does the PIE technique work with non-patterned individuals, such as brown bears?

No, likely facial recognition needs to be applied, such as the approach taken by the BearID project.

Thanks for the feedback. I am part of the BearID Project (I should have mentioned that before). We are looking at alternative models to face recognition in the hopes that we can create an ensemble to improve accuracy further. In anycase, I will take PIE off of my list of papers to read for now!

Ed

Hi Ed,

We'd love to talk about integrating Wildbook with BearID if you have time soon.

Cheers,

Jason

Hi Jason,

Sounds good. Let's set something up. I sent you a direct message with my email address.

Ed

Hi Harold! 

I'd be curious to hear more about this acoustic data logger. Though I have not soldered anything in my life and would probably need an ambulance on stand-by if I attempted it.... But happy to talk about potential applications. My PhD dissertation research incorporates passive acoustic monitoring with ruffed lemurs in the tropical rainforests of Eastern Madagascar (very off-grid, wet, lots of insects and dexterous primates...).

Carly 

Hi Carly,

The neat thing about soldering is that if you suck at it you can try your hand at woodburning instead!

The thing about the soldering is that while much of the electronics assembly can be done cheaply by machine, there are a few pesky connectors that need to be soldered on manually.  My thinking is that if this can be done by the user then a fair chunk of change can be saved.  IMO there are other advantages if the user can be convinced to be more hands on, such as troubleshooting and customisation, like reprogramming the electronics for a completely different mission instead of having to buy new kit.  Field conservationists I have found to be practical people, so perhaps I'm not too far off the mark.

The target application is eventually gunshot/poaching detection, but regular passive acoustic monitoring is a necessary stop along the way.  Since the software is a work in progress there's some opportunity to influence its direction.  For instance apart from scheduled recording I'm looking into recording triggered by certain other events, but I'm open to suggestions especially for niche applications.

I've been concentrating on the electronics and software and haven't considered the enclosure much since this would be very much context specific.  In extreme cases you can resort to a steel pipe enclosure.  I'm not sure why I haven't seen this approach taken more often.  It would take a gorilla with a monkey wrench to dismantle one of those.

Thanks, -harold

Hi Harold, I agree with you, it is important to encourage people to get hands on, that is why I post all the things on my site. Regarding the case, Today I posted two articles about openScad for 3D Design, maybe you find them useful, you can find them here: www.wildedge.info/blog. I started to post about tinyML on the Spresense, besides which board you use, you could find the concepts useful too, I have posts on gunshot detection on my queue. I hope to publish them soon.

What I learnt about soldering:

1)- You need a good soldering Iron and allow it to get to the proper temperature

2)- Pre-warm what you want to get soldered, just a few touches with the hot soldering iron.

3)- Put the soldering iron in one side and the solder in the oposite side, both in contact with the terminal or pin or cable or whatever

4)- when the solder starts to melt, DO NOT PANIC, keep pressing a bit more, then remove the solder but keep a little bit more the soldering iron and then remove it. If you achieve a brilliant silver cone, then you did it right. If it looks like a little ball it will fail sooner than later. To avoid ruining things, practice by soldering cables first.

5)- Use something to support the parts, it is important that they don't move while soldering.

6)- Do it in a well ventilated place and DO NOT BREATH the smoke, that is really dangerous!

Hi! My heart skipped a beat when I read your post. I'm working on data loggers as well for implant vhf trackers in Marmots. I'm using arduino and a radio module to hear the beeps and track the time stamps. We don't have GPS data included as the implants are not small enough. But I'm planning on building a large number of trackers and keep the listening distance down low to "roughly" track the locations. Feel free to bounce your challenges off me.

Hi Ryan,

Interesting stuff!  I imagine your loggers are very small, what parameters do they log?  There's a thread around here somewhere about building VHF tags, an area I've been meaning to experiment with.  I've toyed with the idea of using 433 MHz hardware since they're easy to find and cheap, but they're not natively very small but may be suitable for marmots.

Do you have a writeup of your work?  I'm curious about how you've solved these problems.

Hi,

We're in the 148 to 150 mhz band in VHF. I'd love to have the tags switched over to 433 as they have so many benefits but we have what we have haha. The only real parameter is the time distance between the beeps from a tag. A fast beep means the animal is alive and a slow beep means mortality. Tracking the beeps during the winter while they're in hibernation would be amazing. The biggest challenge I have right now is designing the tracker when we don't have access to one of the tags. They're expensive and Covid is keeping me away from the people that have the tags. The whole project is on hold at the moment sadly while we figure out the logistics. I'm tempted to just order a tag so I can get the data logger working 100%.