Inspiration
We decided to build such an app because we noticed that many people don't understand which waste items go into which bin. This made us want to build an application so that people would be informed on what to do in the future.
What it does
The app asks the user to point the camera at an object. The app will then scan the object and identify it to the closest guess. It will then identify whether the object should go in the compost, waste, or the recycle bin.
How I built it
To identify the object, we used object detection. The app will search through a library of objects, and at the end, pick one object that most closely resembles the object in front of the camera. After the object is identified, the app will try to find whether the object must go into recycling, waste, or compost. It will display its results onto the screen.
Challenges I ran into
Some challenges we ran into really did stump us as there was no way to go around, we tried three different programming languages until Javascript worked. Additionally, we had to use four different types of image recognition until we went with Tensorflow Lite because of its light and stable software. We also ran into the problem of the app crashing because it was under heavy load, we fixed this by unloading unnecessary statements that were creating errors.
Another challenge we ran into was creating the piece of code for the "type of waste". We couldn't figure out what type of an array to use when creating the "type of waste" formula.
Accomplishments that I'm proud of
My partner and I are proud to be able to make a real-life application using Java and Javascript that uses machine learning and object identification. We both are proud to make a code that will help many people in protecting the planet, as without it, we would cease to exist.
What I learned
My partner and I learned a lot about how Java and Javascript work. We got the chance to code for a real-life application on Android Studios for our first time. We also learned a lot about how to connect with other people and comfortably share your idea at a hackathon with other students from different High Schools.
What's next for RoGreen
RoGreen is a small idea developed by two high school students at a hackathon. This small idea has the ability to expand and grow enough to be given other real-life applications.
Built With
- android-studio
- java
- javascript
- json
- object-identification
- tensorflow-lite
- tensorflowlite
Log in or sign up for Devpost to join the conversation.