Inspiration
to reduce carbon footprint by optimising the recycling supply chain decisions
What it does
Detect quality issue ahead, providing better estimates of material composition , forecast delivery time to enable better planning of customers
How we built it
Camera units that can be fitted to containers - send images periodically to cloud storage as the container gets filled Built a raspberry pi prototype Image Classification: Algorithms hosted in the cloud to classify images. Data sent to cloud database. Built 2 prototype algorithms using Azure cognitive services Identify material groups in images Determine how full a container is Mobile App: for staff to view images of consignments and feed info back Forecasting algo and webapp: to view forecasts of estimated delivery of consignments from suppliers. Built prototype
Challenges we ran into
- low manpower.only two team members.
- issue getting the AWS setup working
- issue gettingthe apis working ## Accomplishments that we're proud of
- proved the concept- the machine learning prototypes work, camera unit works, forecasting app works ## What we learned
- bit about recycling , refreshed azure cognitive services skills, improved python skills ## What's next for Recycling optimisation
- it is a scalable solutions and the ideas/components can be used across different companies and industries
Log in or sign up for Devpost to join the conversation.