Limpid

Inspiration

The lack of accessible hardware for detecting microplastics in my own water and wanting to create an easy solution.

What it does

This box helps to control the lighting environment of the water and gives a consistent focal point and angle so that in the future we can use our data to train an ML model.

How we built it

We built it with some 3d design work than was translated via a printer to the physical where we used some salved TV diffusers to get an equal spread of light in the box so we can capture the clearest image. We also used the Super Learning Kits with additional screws salved from a broken TV to secure the pieces in place and add a motor to stir the water around so we can collect multiple images of the same sample to average it out and make larger statements about the body it came from.

Challenges we ran into

So many where to even begin, time was our biggest enemy here from designing the print to the actual task completing was a massive time sink though in that time we did our best to work on preliminary drafts of our code and create the website and landing page that the device would eventually work with to map out all of its datapoints. We also encountered issues with trying to check the math of what we were doing before it was determined that the specs we had been given about focal length and distance from image were wrong and leading to our errors (optics are hard it turns out!). We also had a run in with a half dead bread board but we worked around all of these to come up with a final product we are pretty proud of but the road there wasn't easy.

Accomplishments that we're proud of

We're really proud of our website design and integration with AWS so that we can get individual user data and locations to create a wider data map and the incorporation of Scripps and NOAA to give speed and concentration data of the microplastics to the user if they take an ocean sample. We're also pretty proud of using the iphone camera as an informal microscope as it's rather easy to do, is accessible, and provides remarkable clarity.

What we learned

We learned a lot, the difficulty that comes with optics, the timescales needed for going from design sketches to a working 3d printed model, how to create an AWS user authentication infrastructure to help keep our data clean. There was also the datasets we learned to manipulate and integrate with our own information and how to work without full coverage of an area to give a best guess result!

What's next for Limpid

We hope to gather and create a microplastic in water training set to create and refine an ML model that will be able to identify those particulates current sets out there have a different background than what we created. We'd also like to expand the coastal data we have past San Diego but current information sadly doesn't span that far.

Built With

• amazon-web-services
• noaa-marine-microplastics-data
• scripps-institute-of-oceanography-spray-data