EcoPals

• 

  Main Page

• 

  Scanner

• 

  EcoPal Turtle

Inspiration Shopping centres and town/city centres had the lowest rates of frequent use of recycling facilities, with only one in five (20%) UK citizens reporting they recycled using these facilities “very often”, we hoped to develop a system to help combat the low percentage and imporve city recycling in a fun way.

What it does Our system uses a hand-trained internal real-time machine learning model to detect which recyclable material it can see. Additionally, pressing your ecopal can also reveal local recycling centres with the ability to get directions.

How we built it We found a nano ML model called yolov8 and retrained it using a free to use online database called TrashNet. The front ends are react for the web frontend and react native for the mobile frontend.

Challenges we ran into We faced big challenges with manual training of the model as this is a very time consuming process, similarly, we did not plan the react native app development very well so it is a big more rushed with non-refined UI.

Accomplishments that we're proud of We are very proud of our work with the model - whlst only a proof of concept currently, the systems used were new to the entire team, so to answer what we learned as well, we worked very hard on learning these systems and how to implement then in our own way.

What's next for EcoPals Refining the mobile app UI and greater ML model training would take EcoPals to a whole new level.

Here is our link to a video about our time at the hackathon as well as what the hackathon has taught us and how we have felt during the event. https://discord.com/channels/@me/1127358542974828605/1469903896154603612

Built With

• docker
• git
• github
• node.js
• onnx
• react
• react-native
• typescript
• yolov8