Hi Zoe,

This looks really interesting! Is the initiative on social media at all? I would like to share it with the Australian Citizen Science Association membership. Also, is there involvement in Australia? If on twitter, feel free to connect with me via @JessieLOliver

Cheers,

Jess

Hi Paul,

I recommend you have a look at SciStarter and filter for online projects.

Also here in Australia we have:

DigiVol

Wildlife spotter

Galaxy Explorer

Weather detectives

Explore the sea floor

Trove

and others that are popping up all the time!

Hi Paul, 

I would be interested to see how your project turned out if there is a link to it anywhere? Also, in case it's of interest, UNEP and TED Ed developed the Earth School initiative, and one of the 30 interactive quests focused on citizen science (#26)! The citizen science quest, which I led the development of, included projects that could be done from home by design, being developed in response to COVID. https://ed.ted.com/earth-school

Hi Doug. 

We often work in remote areas like rainforests, desert, Australian outback, etc so we often have to develop custom hardware. We haven't seen anything that fits our needs in a budget range that could accomodate the projects that we work on. 

We've since gottenr requests to make the devices publicly available but are still working on finding time to document and set up the manufacturing for them. 

However you may be interested in one of the designs we've been using in the Middle East and  Australian outback. It's for 3G cellular communications and has inputs for a solar panel and rechargeable batteries. We're manly using it to connect periodically to a cell tower and upload data it collects from other wireless devces (LoRa) in the area. But you can also use it for streaming if your solar/battery capacity is large enough. 

If you're interested us, please direct message us on WildLABS and we can discuss more and perhaps set you up for pre-release hardware. I've included some pics below:

Akiba

Hi Doug. 

It's actually an interesting application, however it might actually be cheaper to try and put together an off-the-shelf solution rather than go custom. There seem to be some livestreaming trailcams that might give you what you want almost out of the box. The Spartan GoLive trailcam seems like it might do this. Also other animal webcam setups seem to use a combination of WiFi security camera + a WiFi-LTE modem. This might actually be the cheaper and quicker way to go.

I do think that having a cellular streaming camera setup specifically for wildlife is a really interesting design. Trailcams are generally designed for camera trap functionality so I suspect the live functionality would be an afterthought. I'm curious if they have to deal with things like heat and recovering the stream after the communications goes down. 

In our designs, we assume a very resource constrained environment, ie: no power sources, little or no connectivity, etc. So looking at a design where it's specifically expecting a power source (ie: dedicated large solar panel) and good/decent cellular connectivity with the design focus on uptime might be something interesting to pursue. 

But for now, I recommend checking out the live trailcams or a wifi security cam + LTE bridge setup. And if you're interested in a dedicated solution specifically for wildlife cams, I'd love to discuss what the requirements would be. It might be super useful for wildlife orgs and parks to build awareness and improve funding,as well as benefit the research and conservation community.

Akiba  

That's such a cool list. Thanks for posting it!

Akiba

Oh, so excited to hear more about what's been going on. Looking forward to watching Tech Tutors 3 also. There's been something missing in my late evenings :)

A suggestion for something I repeatedly get asked:

- How do I create maps of land cover change for my area and assess accuracy, using free software and limited (or no) coding skills?

Any takers?? I keep meaning to take a look at the latest plugins in QGIS, for example, but haven't got around to it...

Ollie

Hi Wilson, Similar product we are creating at https://aivi.in . We are focused on audio, image, vedio based intelligent and based out of Bangalore,India. Currently implementing this solution( Computer vision, IoT , Cloud , Intelligent) for one of santuary in India to detect poachers and animals

Attaching idea snap that we are implementing. Intellgence Dashboard provides all related details ( Count etc.)  This is just driven by IoT enabled camera and Mobile App for plug and play solution

I'd highly suggest going with a pre-built service like AWS Rekognition or Google Vision, as their models are already trained with more images that you could possible do yourself.

hi Louise, I don't know if you are still looking, but we've just become second (5 digits behind the comma ;-) at the iWildcam2021 Kaggle competition. We're currently building a full-circle solution to:

• identify objects, people and animals on images
• determine whether they form a time sequence (several images in a row showing the same animal that passes by)
• count the number of object/animals seen in that sequence
• place classifications that do not pass a user-defined threshhold in a queue
• provide an interface to manually classify these images
• and pass back the results to a client. This could be an API, or our any of the analytical tools in out WITS platform

If you are interested in a call, you can reach me at [email protected],

Best regards!
Jan Kees

Hi,

Maybe try sourcing from Aliexpress?  I an't speak to the quality, but at some price point it becomes worth looking into this option.

-harold

Hi Antoine,

I have worked extensively with livestock guarding dogs (Anatolians) in the past. The effectiveness of this conflict mitigation method depends largely on how well the dog and the livestock bond to each other. It makes use of the pack instinct of the dogs, bonding to the livestock as his pack to defend from attackers (including other predators).

So the first question would be what typical dog behaviour do you want to exploit for your "crop protection dogs"? I would assume that you want to use territorial behaviour as the basis of the dogs protecting the croplands as "their" territory. So your training should focus on this (not sure how you would do it, though). The second issue could be that the dogs might not recognize elephants as intruders into their territory; i.e. they might be come used to the presence of crop-raiding elephants over time -- I really do not know dog-elephant interactions well enough to make a prediction on how this might play out.

A third  potential issue that you need to think about, is exactly how effective the barking dogs will be as deterrents. Even if they warn the local farmers, will the farmers be able to drive the elephants away without danger to themselves? And will the dogs be effective in deterring the elephants, or rather be an irritant that makes the elephants (more) aggressive? I don't know. Will the dogs actually attack the elephants, or be scared of them (and maybe run away towards the village while being chased by the elephants)?

I think if you can address all of these points effectively, there is a good chance that this approach will succeed.

Cheers,

Chavoux