TrennMich

• 

  icon

• 

  Software Architecture

Inspiration

We ourselves sometimes have a hard time by separating waste, correctly. Especially, when you are moving to another place you have to watch out for the constraints of the local community. It is a fact that seperating waste in a prescribed way helps the environment and every single person can be part of that. Progress is progress, no matter how small.

What it does

The main objective of our project is to achieve the best possible separation of waste. For this reason, we offer a smartphone app with a good overview and easy handling. First, users can enter their zip code. They are then redirected to the next page, where they can take a picture with their smartphone camera. Using artificial intelligence, our program recognizes the current object and assigns it to the correct waste separation. If you have to go to a recycling center, the nearest ones are displayed.

How we built it

The backend consists of a Java Spring Boot application which stores the data entities in a Postgres instance. It provides an REST api to the react native app.

Challenges we ran into

Including the Pytorch model in the spring boot application turned out to be very challenging. We used Deep Java Library to include the model in the server but ran into some problems. Eventually, we decided to use an own Flask instance for the model, which solved the problem. Another big challenge for us was the framework react native, since none of the team had any previous experience with it.

Accomplishments that we're proud of

Although React Native was new to all of us, we were able to get up to speed and develop a decent app with an appealing user interface.

What we learned

We learned react native and advanced in scraping. We also learned a lot about team cohesion and perseverance.

What's next for TrennMich

More cities and counties with different waste managements can get involved in this project. There will also be the possibility of identifying more objects with the camera. Another milestone that can be achieved is an expansion of this project in terms of barcode scanning on already consumed products to get an optimal separation of waste.

Built With

• java
• postgresql
• python
• pytorch
• react-native
• spring
• typescript