|
|
|
|
|
by TwiztidK
740 days ago
|
|
I'm really curious how this would've looked if he'd taken a mid-range budget approach to vision instead of being able to throw (tens of?) thousands of dollars of industrial motion capture cameras at it. Something like 2-4 global shutter cameras, all running a similar vision filter to what he first demo'd, feeding a kalman filter/state estimator to determine the ball's position. I can't blame hime for following the mantra of "don't build what you can buy (or get from a sponsor)" but it would be cool to see some affordable solutions (or attempted solutions) to this problem. |
|