Collaboration with Hamamatsu + GroupGets on single pixel camera for hyperspectral imaging

hamamatsu
singlepixelcamera
hyperspectral
goshcollaboration

#1

Hi all, so I’ve been pursuing the single pixel camera collaboration, specifically the application for creating a low-cost hyperspectral imager. I have an update.

I talked with Hamamatsu (http://www.hamamatsu.com/) today, and they are interested in supplying development kits to collaborators in the project (yay!)!!! The specific spec we will be working on is: http://www.hamamatsu.com/eu/en/product/category/5001/4016/C12880MA/index.html. It’s 350 - 850nm spec with 15nm resolution.

I also talked with GroupGets (https://groupgets.com/) and they are providing 6 breakout boards so we can hook the specs right up to an arduino or teensy, which will way simplify our development. If you haven’t seen their breakout boards - go check them out: https://groupgets.com/manufacturers/hamamatsu-photonics/products/c12880ma-micro-spectrometer .

The Deal:

  • They would supply arduino compatible development boards (so no soldering necessary) to collaborators (normally a few hundred bucks).
  • In exchange for their support, we would need to publish the results of our efforts (Journal of Open Science Hardware would be ideal), and have an publicly documented project online which uses the Hamamatsu mini spec.

Next Steps:

  • Apply if you’re want to contribute and get a spec by replying to this post noting your interest. Applications are closed as we only had 6 units available in the first round and those were filled. But, in case we get another batch and you want to be considered, please fill out the application form is here: https://goo.gl/forms/x9ue6mmfwIUxY2qk2 . Please only people willing to actually work on this sometime in the next year… so still respond if you’re next month is busy but you can put in some time over the next year. That’s totally ok. Anyone can apply, but useful skills will be mechanical design, optics, math (the algorithm to reconstruct the image), application identification, or those who want to help write/publish the results (people who want pubs are welcome!). If you don’t have those skills but think you can still help, just mention how in your response.
  • Include a description of an application for the hyperspectral imager in your response. This was a requirement of Hamamatsu - it helps justify their investment in us by describing potential markets and applications.

Most of the firmware was already done, so we’d be ahead of the game in terms of development.

Want to cross post this request for collaborators? Yes please do! Just make sure to direct people back to this post to respond to the request, and please repost to places where you think people will be a good fit in terms of skills and talents for this particular project (see above for list of skills). Because Hamamatsu will be supporting us with equipment, I want to be respectful of their generosity and make sure we’re not sending things out willy nilly. So a reasonable skill fit + available time to contribute are the baseline for participating.

Why is this cool? If you’re not familiar, there are many great applications, besides just as an educational tool. https://en.wikipedia.org/wiki/Hyperspectral_imaging#Applications , or here’s a nice article - https://www.linkedin.com/pulse/hyperspectral-imaging-whats-all-why-we-talking-now-laith-marmash .

@dusjagr @gaudi @rbowman @marinappdf @andrew.david.thaler @amchagas and @jeff you are the ones who expressed some interest in this, let me know what you think.


#2

Hey @gbathree Great! Any considerations on cross-posting this offer?


#3

Definitely! Where would you suggest?


#4

Personally-- the general Public Lab list (we just had a similar type of collab happen on the water list), but others on the forum have their own networks as well they might want to send to. Maybe just putting some reusable language together that people can send out and noting either to come back to this forum to comment on the thread if they’re interested or who to connect with directly?


#5

Ok - updated with information about cross posting - does that look ok?


#6

@gbathtree are you happy for this project to be recategorised out of #gosh-2017 and into #general messages, as it’s on-going?

Jenny


#7

@TAKE can maybe pick some up by himself.

i’d be interested to work with it myself. ideally another model that goes higher into the NIR.

m


#8

Wow, that’s sounds nice!
Thank you Marc.

Hi,
I am Take living in Hamamatsu-city.
Hamamatsu’s head office is around 30min from my lab.
And have two friends working in Hamamatsu(not engineer of optical sensor… they are engineer of LASER and PET).
So, I would like to help if you need local support.

Also, I have interest in Agriculture projects.
Making robot, drone…

Take


#9

@dusjagr - I thought you guys might be interested.

@TAKE , that’s fantastic! Having someone close to headquarters would be fantastic

@jcm80, yes, sorry should have posted appropriately the first time :slight_smile:

@shannond I added language for cross posting, i think Public Lab would be a great place. If anyone else wants to cross post please do, just read the language I added to the main post at the top.

Ideally I’d like to find 5 - 10 collaborators.


#10

Hi,

I just signed up for this forum.

My name is Markos, I’m chemist and have been working on projects related to Water.

I am currently starting a project for photometric measurements for water quality.

Most of my projects are in Portuguese on the site:
www.c2o.pro.br/hackaguas/index.html

But I have translated something to English on:
www.c2o.pro.br/en/automation/index.html

And I’m interested in collaborating with this project of Hamamatsu.
How can I apply to participate in this project?

Thanks,
Markos


#11

Hi Greg,

This project looks really interesting!

If it helps, I could contribute in testing/validating/benchmarking the device. I have done something like that comparing vegetation indices from RGB and IR-modified consumer cameras (including publiclab’s infragram) to a (very expensive) hyperspectral camera (I will publish those results in a paper soon). Here’s a picture of my setup:

I did this in Germany (unfortunately I don’t own that camera), but I already contacted my german friends, and it’s quite possible (provided I work out some institutional issues) that I could perform some similar tests with the single pixel camera, in order to compare the performance against a ‘standard’ device (probably late 2017 - early 2018…).

Do you think this could be useful?

And regarding applications:

  • In my lab/institute, we would be interested in analyzing plants, and possibly also food quality (e.g. potatoes, fruits). I don’t know the expected size or weight of the device, but it would also be great for UAV aerial imaging of crops (but taking into account the 20s for each frame… so maybe ‘static’ aerial imaging)

  • Since hyperspectral imaging can be useful for plant phenotyping, I could try to get other people from the IPPN involved in the project (I could present it during the workshop next month)

Saludos!
Gustavo


#12

OK - so I moved this to it’s own community in the GOSH forum (https://forum.openhardware.science/c/communities/single-pixel-camera-dev-group), so we can have discussions and stuff here in a semi-organized way :slight_smile:

@Markos welcome! Here’s the application - https://goo.gl/forms/x9ue6mmfwIUxY2qk2

@gpereyrairujo that would be perfect! We’ll definitely need comparisons against existing equipment, that would be so helpful. Also, the fact that you have specific applications is helpful in showing Hamamatsu that this is worth some investment on their part. Finally, this may be a great opportunity to publish for you as well if that’s of interest.

Seems like things are moving along pretty quick, so I made an application form which will get us what Hamamatsu and GroupGets need to move forward.

So to apply fill out this form! - https://goo.gl/forms/x9ue6mmfwIUxY2qk2

Sorry this is highly undemocratic, but I don’t want to bog this down with process either. We have about 6 units to start, so I’m going to take applications and get a mix of skill and first come first serve, in this first round. If we make good progress, it’s possible we could go back and ask for more.

Also, if people are super uncomfortable with me taking the lead on this or want to do this selection a different way, please let me know, I’m open to options here, I’m just trying to keep up the momentum.