ssoria / flatland-arg Goto Github PK

Now that we're moving the IR target and the microcontroller onto the costume, we need to experiment with how the player expresses the intent to preform a gesture. This is important because we do not want to falsely recognize a gesture when the player is just running around the field. The two options that I know of so far are this:

• Use an onscreen action, such as a double tap on the screen to bring up a menu
• Try recognizing gestures all the time and keep track of previous recognized gesture. Then only register affirmative gesture if previous and current match within a time limit.

need to get the gesture recognition working with the rest of the system

Here's the math that peter came up with for correcting the camera skew from the previous incarnation of the software:
https://github.com/uclagamelab/flatland_cameravision/blob/master/CorrectLocation.pde
https://github.com/uclagamelab/flatland_cameravision/blob/master/Tracking.pde#L225

I'm ready to start integrating the tracking application (which is written in c++). so I was wondering what the best way to integrate this with the rest of the system would be. I could create variant of the Flatland client that opens a Unix socket with the camera tracking app, but then I was thinking that it might be better for the Flatland game server to open an Internet socket connection with the the camera tracking app. What do you think?