Mothbox 4.0
digital-naturalism-laboratories.github.io203 points by bookofjoe 18 hours ago
203 points by bookofjoe 18 hours ago
Hi! I'm andy! I'm the guy who spent the past year making the Mothbox! I don't know how this got on ycombinator, but cool! We are working here together Let me know if you have any questions!
Some quick basics: It's an entirely open source system made from off-the-shelf parts made so you can build one yourself! We built it to help a set of community reforestation initiatives here in Panama! (
Existing automated insect monitors we checked out wouldn't work for us because: 1) They were REALLY BIG (like huge heavy suitcase size+) 2) They were really EXPENSIVE (like 7000-15000$ !) 3) There wasn't any info to build robust ones that could withstand months in the rainforest!
So we build one ourselves! We started with absolutely no budget, and the costs that people have noticed ($375) are really the cheapest you can find these parts (Pi5 + big battery + 64MP camera + really bright lights), and for instance are on par with the $200 just for the 16mp logitech webcam employed in many existing insect monitors!
We got about 20 of these out in the wild here in Panama already, and are literally building more in the jungle as I'm typing this!
Here is how you got on ycombinator:
1. I happened on your website in Clive Thompson's newsletter, among the websites in his final "sudden death" round:
https://buttondown.com/clivethompson/archive/linkfest-25-the...
2. I thought HN peeps would find it of interest so I submitted it
Hey Andy, this looks awesome. Nice work!
I think I'd like to get one to track insect diversity and populations in areas where I work on riparian restoration projects. My hope is that when projects are completed, in the following years it'll be evident in the insect data the Mothbox collects. Do you think that's a valid use case? Maybe I'd need more than one to get enough data?
BioClip (referenced in their writeup) is good to know about.
I run a small scale wildlife monitoring program with usually 3 cameras (depending on how many are being repaired for battery terminal corrosion). Each month I skim through and categorize ~200 videos that we've captured. Doesn't take too long, but it would be nice to automate at some point.
Do you use any kind of protectant, like dielectric grease or Deoxit? I find applying to both the equipment and battery contacts saves a lot of headaches on my work phone, ATV, trailer lights connectors and anything else powered I use in the field.
We use some dielectric grease when we have to make some outdoor connections and it works great here in the jungle, and it also has the bonus of keeping the leaf cutter ants from slicing up your electronics too!
No, but great idea. Once I get the current batch back from the repair shop I'll start doing that.
BioCLIP won the CVPR best student paper award: https://cvpr.thecvf.com/Conferences/2024/News/Awards
Having talked with Sam, they're working on BioCLIPv2. So expect even better results sometime soon.
Are there any "citizen science" initiatives where I can buy a device like this and upload my insect counts to some database that's useful for researchers?
I have been interested in setting up some things like home weather monitoring, ADS-B, streaming webcams for wildlife, etc. anyway so this would be a fun item to add to that.
Cornell is doing something along those lines for birds using audio recordings. It’s just a smartphone app though - https://merlin.allaboutbirds.org/sound-id/
I was using it today, and any day when there are birds around. It correctly identifies birds you can barely hear. It's become as useful as binoculars to me. A great free app.
I have a pi running birdnet-pi on my porch. https://www.birdweather.com/birdnetpi
they also release the model. I've recorded hours of audio outside my house and run it through the model on my desktop with GPU.
I think the air gradient folks do some collaboration and it's possible you could get your sensors linked into a bigger net by contacting them: https://www.airgradient.com/research/
Edit: Also MIT did this recently: https://news.mit.edu/2023/low-cost-device-can-measure-air-po...
On a related note, I wish there was some sort of software system where I could ID weeds and insects with photos and locations at various levels of confidence (“possibly”, “probably”, keyed out, genetic testing) at different taxonomic levels and both self host my results while also automatically contributing to a larger project.
There are various projects out there (like plantnet) but I don’t want to burn massive personal effort curating in a system that isn’t my own first and foremost, due to inevitable enshittification. At the same time, I want others to benefit from what I do, in particular local growers and naturalists. Things like PlantNet also tend to be “majority vote” on ID, meaning a whole lot is often close, but wrong. For example there is a regional plant specific to my area called Willamette Navarretia. Those that don’t realize this will easily confuse it with very similar looking species found most elsewhere in the western US, and last I checked it wasn’t in PlantNet.
https://www.oregon.gov/oda/programs/PlantConservation/SiteAs...
https://www.inaturalist.org/taxa/165663-Navarretia-willamett...
If you want to manage your own data take a look at TaxonWorks. Multi-entry and bifurcating key editors, images, much more. It, and its companions are all open source, see TaxonPages for example.
Have you tried iNaturalist? [0]
Insects, animals, plants, etc get posted from all over the world for ID.
[0] https://www.inaturalist.org/
I post lots of photos of things I find locally and experts step in and help ID the subjects when I don't know what I'm seeing.
I also have a couple of AudioMoths for recording local sounds including birds, insects, etc. Very high quality units at a reasonable price.
I have seen this Mothbox posted a couple of times and almost bought one since I know I have tons of moths attracted to my native plants out here, but the price is a sticking point right now. I think something like this combined with an AudioMoth and some trail cams would be a good local wildlife monitor setup.
edit so it turns out, at least a significant part of inaturalist is in fact open source of some kind. I never realized, I’ll have to look into this more. I appreciate you getting me to take a second look https://github.com/inaturalist
This may be exactly what I need: https://www.inaturalist.org/sites/network
> due to inevitable enshittification
PlantNet is led by an alphabet soup consortium of French research institutes like IRD, CIRAD, INRIA, and INRA so I don’t think enshitification is inevitable - though I don’t know why it’s not open source. It’s funded by grants and donations with no incentive to enshitify it for more revenue.
This could be a great teaching tool, BUT at about $375 per Mothbox, it's going to be a hard sell for most schools.
I don’t think it’ll be a hard sell for schools at that price, school budgets tend to be fairly elastic for new things (and sometimes for wholly unnecessary things, for the price of less than 1/4th of new MacBook for the principal (for example) they could have this.
Edit: also if schools were to be interested in this (which they should be, it’s very neat) they could group together and buy in bulk which would also greatly reduce the price
$375 per box doesn't seem bad to me at all when you probably only need a couple per school?
Yeah, i feel like currently they are at about the price of camera traps 10 years ago. There is very little mass-manufacturability to them right now (it's all open source and made from off-the-shelf parts) but later if we can find more funding, we are going to make a design more for manufacturing which should hopefully drive the costs down even more! :)
> There is very little mass-manufacturability to them right now (it's all open source and made from off-the-shelf parts)
This is the obstruction to using them in an educational setting. If they were available for $600+ each but already completely built (minimal DIY), they would be more likely to get into (some) schools.
OTOH, it'd a fun science project just to built one, for maybe a different set of kids that operate the box.
Just needs motivated teachers, if you ask me. I assume the mothbox is more of a high-school project, building one seems on that level as well.