Multi-Display Data Visualization
Data Visualization on a 22 Display Raspberry Pi cluster
I created a series of projects, some relating to data visualization, others not, on a multi-display rpi cluster. This page contains some of those projects. For the most part, projects used the python Message Passing Interface Library mpi to synchronize rendering.
VID_20181202_173145.mp4Milky Way
15 Display rotating visualization of all of the known starts in the Milky Way (as of November 2018). Coordinates are converted into Cartesian space so they can be rendered on the screen.
aruco_data.mp4Data Interaction
Using ArUco markers and OpenCV to get the 3D pose of cubes. Relative poses were then used to rotate a 3D data visualization.
Final implementation of this visual was done on a 9 panel display (3x3) with the additional ability to select points on the graphs for more information
Testing an interactive gallery projection:
VID_20181128_001526.mp4Interactive Sidewalk
The left video is a prototype of a project which used OpenCV to detect people walking along a sidewalk in front of a gallery (in which I was showing work). As people walked along the sidewalk, a projector would render a particle system of little yellow triangles falling towards the gallery space.
Sadly, this wasn't displayed or documented as the projector I was using wasn't bright enough to outshine the street lights nearby.
I also briefly did some work using Vuforia markers on multi-panel displays to control the display of AR content. I didn't have time to pursue this project very far, but the goal was to create a markup language for AR using markers. A user could walk up to a display panel and have AR content populate the space based on those markers - sliders, buttons, text - the standard HTML elements.
VID_20181209_173347.mp4Video taken from the Exhibition
(the bubble wrap project is not mine)