CVBasketball

• 

  The HSV mask created to determine the green marker.

• 

  The ball mid-air in the court.

Inspiration

We realized that there is a lack of computer vision in many important sectors of life. One that we recognized as significantly lacking was sports. We looked into the basketball specifically, and how we could use color detection to receive the position of a basketball, and predict where it would land in real life.

What it does

We analyze colors that we see through the camera, specifically the green hand indicator that we created, and use custom conversions to place the ball in real life dimensions. We then use physics and various algorithms to display the ball in a graphics window.

How we built it

We used Python to create the CV algorithms and Processing window.

Challenges we ran into

Our physics model did not work exactly the way we wanted, primarily due to the lack of the ability to precisely determine the velocity of the ball once it launches. We also had issues with the path of the ball once we launched it.

Accomplishments that we're proud of

We successfully created an accurate CV algorithm to detect colors and correlate the real/camera position into a processing window. We also were able to get quite far with the physics, but not enough to properly demo.

What we learned

We learned about various processing libraries for python as well as how to use running averages to properly determine the moment the ball is released.

What's next for CVBasketball

We will continue to work on the physics behind the project and update the GUI to make it look better.

Built With

• imutils
• numpy
• opencv
• p5
• processing
• python