Inspiration
AR can be broadly split into two categories -- marker-based vs markerless. One way the latter can be achieved is by constructing a 3D map of the surroundings from a robot's visual input while in motion, through methods such as SLAM (Simultaneous localization and mapping) or more recently, DSO (Direct Sparse Odometry).
While interfacing the GPL-licensed DSO implementation with a live camera feed was our original goal, we have instead resorted to... "Markerless" AR.
What it does
Superimposes a spinning Boston Hacks medallion on "markerless" markers within the camera's viewing frustrum.
How we built it
Unity, Vuforia
Challenges we ran into
- Building project on OS X was a pain, used an AWS EC2 instance instead
- AWS EC2 instance wasn't GPU-enabled for OpenGL, had to relaunch with a g2 instance
- X11 forwarding is slightly broken with NVidia drivers because of an issue with XQuartz (or so John says)
- Tried using VirtualBox to build: Virtual memory exhausted, VirtualBox version errors because apparently vendors don't comply with a fixed standard for .ova files??
- DSO code is a lot to take in without proper documentation
- Cordova Unity plugin was too limited for our usage.
Accomplishments that we're proud of
- Building the project on VirtualBox. Finally.
- Gaining some semblance of understanding of how the DSO project works
- Being trolls
What we learned
- Sleep is important
What's next for Markerless AR
- To accommodate Markerless's of all shapes and sizes.
Log in or sign up for Devpost to join the conversation.