Inspiration

Every day, people unfortunately litter. To avoid adverse effects on the environment, cities have a variety of environmental cleanup programs to alleviate this problem. As volunteers work towards cleaning up the city, however, they often end up spending a great deal of time also cataloguing the types of garbage that they see throughout their days. Later on, they report back to the city program managers who then use this data to categorize littered waste and determine the economic impact of the littered waste, as well as to organize educational campaigns around specific categories of waste that the city seems to be accumulating the most of outdoors. This manual method of tracking waste is time consuming for volunteers and employees, so we built a more automated, robust method of tracking and categorizing littered waste using Microsoft Cognitive Services and Microsoft Azure, called ECO Request.

ECO Request allows users (cleanup volunteers and employees) to take photos of waste and upload them as an "issue" that gets sent to the city or whoever is leading the environmental cleanup efforts. Through this process, Cognitive Services will categorize the type of waste the cleanup individual processed, adding it into a database that aggregates all of the waste processed, photographed, and sent through ECO Request, into a list of pre-determined categories and tags of waste types (plastic, bottles, cans, etc.). In doing so, cities will gather more robust data on the kinds of waste they accrue, and it will allow them to better educate the public in the future, as well as better handle cleaning up and disposal of specific types of waste they accumulate most often.

What it does

The public Android app allows users to submit pictures of garbage in the city. The city worker management app shows all the pictures submitted by the public users. The backend uses Microsoft Cognitive Services to categorize the garbage type.

How we built it

  • Android app - Android Studio
  • City Worker App - Javascript/Bootstrap
  • Backend - Node.js, MongoDB on Azure
  • Computer Vision - Microsoft Cognitive Services

Challenges we ran into

Image categorization

Accomplishments that we're proud of

  • Easy to use public app and city worker app
  • Categorization of garbage images

What we learned

  • Azure
  • Microsoft Cognitive Services Computer Vision

What's next for ECO Request

  • Improving image categorization
  • Improving public app
Share this project:

Updates