We saw how Pokémon Go inspired a ton of people to walk and travel way more than they did before. We thought we could design an app to incentivize a culture of sustainability.
We also feel that this application could be used by non-profit organizations or volunteers on finding the most polluted areas in their community.
We could even find the most optimal location for trash cans!
What it does
Load the website on any internet device (not deployed so picture taking works only on local machines but phones can access the site), take a photo of a spot of trash, and get points for reporting it. Find a spot of trash near you on the website, take a photo of it, and get even more points for clearing it!
By using the google cloud computer vision API, we can gather information of the trash that is contained within an image. We can then give you points based on how polluted the location is.
How we built it
We used Vue.js for the front-end, Django for the back-end, and Google Cloud to perform image recognition tasks and handled the PostgreSQL SQL. Web browser APIs help manage location data.
Challenges we ran into
Getting an image from the front-end web app, encoding it and sending it to the back-end and then feeding that to the google API took a very long time a lots of testing.
There is also very limited functionality on capturing multiple objects from one static image and specifying what objects we are looking for in an image (to our understanding). Regardless the machine vision API was immensely helpful.
We also ran into challenges finding the best database for our project. We ran local ones in docker images which crashed and had problems using some cloud providers. We then settled on Google's SQL engine as it gave us a nice interface and made it very easy to access the database instance for developing.
Accomplishments that we're proud of
Being able to present a easy to use interface for taking photos and passing the picture across multiple back-ends and APIs in a process that seems near real-time to the user. While we can't do it on the phone we feel that we accomplished quite a bit.
What we learned
Lots of Web Programming!
What's next for Where Da Trash At
Getting the app to work on the phone Filtering points by geolocation- such that you can get points in say a .5 mile radius or any arbitrary distance a user wants Make it look even better.