-
-
Image showing Apriltags used to bound the whiteboard writing area. All 4 tags can be printed on one sheet of A4 paper.
-
Whiteboard taken at an angle with a clipboard in the center for reference.
-
Adjusted writing area from the previous picture.
Inspiration
We were inspired by a realization that communicating visual ideas remotely is very difficult without drawing enabled devices. These devices can be quite expensive and niche so many people don't have them. Our project aims to solve this problem using AR to create a virtual whiteboard that people can use to communicate visual ideas, make presentations, and work together more effectively in quarantine.
What it does
Our program uses computer vision to align the virtual whiteboard on a real world surface, it then communicates with its server hosted on Google Cloud to sync the current diagram with other users and performs a perspective transform on that diagram to make it look as if it were placed in the real world.
How we built it
Python, OpenCV, and Apriltags
Challenges we ran into
Learning the necessary OpenCV computer vision functions in order to accomplish our job.
Accomplishments that we're proud of
It works very well for projecting a 2d image into the real world.
What we learned
Using OpenCV to do perspective projections as well as hand detection using Haar cascade machine learning.
What's next for Virtual Whiteboard
Better Hand Detection Tagless 3d pose detection Better networking
Built With
- apriltag
- google-cloud
- opencv
- python
Log in or sign up for Devpost to join the conversation.