Inspiration
Every year we dump a massive 2.12 billion tons of waste. To put this into another perspective, if all these waste was put on trucks they would go around the world 24 times. This amount of waste is partly because most of the things we buy is actually trashed within 6 months.
Inspired to reduce waste emission, we created Waste Advisor.
What it does
Waste Advisor is a computer vision based app that analyzes the things you throw to the trash with easy and simple steps.
You just need to upload the photos of your trash to our web app. Wait for a short while between each image upload and clicking view result, which will help you get to know with your waste! Our computer vision based app would know what kind of waste you produce and gives you feedback based on it.
We hope that this app would help individuals make more sustainable choices to reduce the waste they produce, which would require less non-reusable production in the future thus reducing carbon emissions.
How we built it
We built our front end with react and typescript that connects to Azure storage account for storing uploaded photos, and we built our backend using Java which deals with processing the results from Azure computer vision and custom vision prediction model. We used Heroku for deploying our project.
Challenges we ran into
- Setting up the storage account access from the front end web app
- Training a new computer vision model with limited inputs of images
- New features and restrictions brought by TypeScript
Accomplishments that we're proud of
This is our first project based on a cloud computering system, we are excited to see the possibilities it opens up for our project.
What we learned
The things we learned in DubHacks has been extremely valuable. We learned new languages, APIs, frameworks and how to use cloud services. We also learned that communication is key to collaborative work.
What's next for Waste Advisor
- We could get a better prediction model by finding and tagging more trash categories.
- We could add features that gives a more precise feedback to the user, by suggesting related online articles, and near-by stores that have sustainable practices.
- We could develop hardware (cameras and sensors) to automate the image upload process (when the sensor senses the garbage can opens, the camera will take a few photos) to make it more convenient for users.
Log in or sign up for Devpost to join the conversation.