CoinCounter

CoinCounter is the new way of checking out. It offers a method of counting coins that is fast and mitigates germ and COVID 19 spread using just an image from your pheon camera!

Comment

Inspiration

When recently packing to move out of my home town and to Vancouver, I stumbled across an ancient container filled with my old coins. Being in a rush, I didn’t count the coins in there and just threw the container off to the side. An app such as CoinCounter would’ve been immensely useful then as I could’ve gotten the total value of those coins in a matter of seconds.

What it does

Coin counter takes in an image and uses the OpenCV Java library to determine each coin denomination and find the total dollar value of the coins in the image.

How I built it

The image is preprocessed using Greyscale, Gausian Blur, and Conny Edge detection schemes before being given to the OpenCV HoughCircles algorithm. the HoughCircles algorithm returns a list of the circles in the image along with their radiuses and positions. The circles radiuses are then passed to our Denomination module to determine what coins are what and their respective values using the specific size ratios of Canadian currency.

Challenges I ran into

  • Coming up with a robust and consistent pre-processing scheme.
  • Environment and Dependency Challenges.

Accomplishments that I'm proud of

  • Completing an working prototype for our CoinCounter project.
  • Gaining experience with Maven and Git.
  • Understanding OpenCV in Java.
  • Managing a team under tight time constraints.
  • Working and learning greatly as a team!

What I learned

Being my first remote hackathon, nwHacks gave me a unique opportunity to work alongside a fully remote team. I gained both hard and soft skills, and hope to leverage them in future hackathons and jobs. In terms of technical skills, I gained the most experience using more functionality of Git. Branches and pull requests were new to me 24 hours ago, but now I feel comfortable with them. My communication skills took a big step forward as I communicated with my teammates using Zoom, Discord, and commit messages.

What's next for CoinCounter

  • Improve our interface to display all denominations of coins and their quantities.
  • Improve the reliability of our coin detection algorithm using Machine learning neural nets to detect coin faces.

Built With

Share this project:

Updates