loading
 
Vision based commands
Dave F from United States  [3 posts]
12 year
Is there a way to use the shape matching feature to command a robot (in my case a Vex Protobot) to do something based on the input?  For instance, if I wanted the robot to, say, turn 90 degrees counterclockwise when it sees a square (an image that would already be known to the robot), could I, and how could I, do that?  I'm new to the program, and for all I know the answer is blatantly obvious, but I would greatly appreciate any help.
EDV  [328 posts] 12 year
It seems that you are looking for something like this:
http://www.youtube.com/watch?v=kZKr70A759Y
http://www.youtube.com/watch?v=HTxNlOpm11U

Here are some examples of simple programs:
http://forums.trossenrobotics.com/showthread.php?4764-Using-of-AVM-plugin-in-RoboRealm&p=48865#post48865
Anonymous 12 year
Dave,

As EDV points out using Fiducials or Glyphs is what you probably want to use. You can also have a look at the Fiducial module

http://www.roborealm.com/help/Fiducial.php

and react based on the name of the image being recognized. The advantages of using fiducials is that they can be recognized at many orientations and different angles which is normally the case if someone is showing the shape/fiducial and would not do so in a perfectly perpendicular angle.

STeven.
Dave F from United States  [3 posts] 12 year
Thanks.  That's exactly what I need- the one thing I need to know is how to get the input from the webcam into the microcontroller ( I plan on using a wireless webcam) so it can be compared to the fiducial database.
Anonymous 12 year
Dave,

Just a clarification .. the fiducial database would be on a PC instead of the MCU ... which most likely may not have enough room to keep that on it anyhow. Typically the sequence is Camera -> PC -> MCU with the PC doing the vision processing and sending a final move command to the MCU over something like serial.

If you really want to remove the PC from the mix you'd be looking at something like a BeagleBone board which is capable of doing some vision processing but you'd then be programming that using a much lower level system than RR. RR does not yet run on the BeagleBone board.

Naturally there are many other boards then the one mentioned above that can do a similar task ... that one just happened to be in mind.

STeven.

This forum thread has been closed due to inactivity (more than 4 months) or number of replies (more than 50 messages). Please start a New Post and enter a new forum thread with the appropriate title.

 New Post   Forum Index