by PedroR » Mon Jul 13, 2009 12:56 pm
by PedroR
Mon Jul 13, 2009 12:56 pm
Hi sniperscope
Good to know about your project
With regards to sensors/cam/sound my recommendation is this:
- Use a USB webcam on the Roboard and use some Open source vision library.
I recommend this
http://opencv.willowgarage.com/wiki/#We ... sOpenCV.3F . Form the website these are the supported capabilities "Object Identification, Segmentation and Recognition; Face Recognition; Gesture Recognition; Motion Tracking, Ego Motion, Motion Understanding; Structure From Motion (SFM); Stereo and Multi-Camera Calibration and Depth Computation; Mobile Robotics."
- For sensors I would get a set of Sharp IR distance sensors.
I would maybe assemble them on top of a servo so you can "scan" the area left to right or top to bottom and avoid blind spots as much as possible.
- For sound, since the Roboard has sound output and sound input, I think you can get a simple speaker and an omnidirectional microphone.
Back in the late 90's Microsoft Research was giving away some speech recognition libraries but I believe those are not available anymore.
It was quite simple to add some code there to have the computer recognize and react to simple one-word or two-word commands.
Running vision + speech recognition in real time may be a bit too much of a stretch for the Roboard but running them in turns may work.
I would leave Speech recognition at the end of the priority list though and invest in the vision library
If you run into any roboard issue when assembling everything, please post on our Roboard forum as well, because the Roboard engineers keep monitoring the forum and you'll certainly get good info there as well.
Regards
Pedro
Hi sniperscope
Good to know about your project
With regards to sensors/cam/sound my recommendation is this:
- Use a USB webcam on the Roboard and use some Open source vision library.
I recommend this
http://opencv.willowgarage.com/wiki/#We ... sOpenCV.3F . Form the website these are the supported capabilities "Object Identification, Segmentation and Recognition; Face Recognition; Gesture Recognition; Motion Tracking, Ego Motion, Motion Understanding; Structure From Motion (SFM); Stereo and Multi-Camera Calibration and Depth Computation; Mobile Robotics."
- For sensors I would get a set of Sharp IR distance sensors.
I would maybe assemble them on top of a servo so you can "scan" the area left to right or top to bottom and avoid blind spots as much as possible.
- For sound, since the Roboard has sound output and sound input, I think you can get a simple speaker and an omnidirectional microphone.
Back in the late 90's Microsoft Research was giving away some speech recognition libraries but I believe those are not available anymore.
It was quite simple to add some code there to have the computer recognize and react to simple one-word or two-word commands.
Running vision + speech recognition in real time may be a bit too much of a stretch for the Roboard but running them in turns may work.
I would leave Speech recognition at the end of the priority list though and invest in the vision library
If you run into any roboard issue when assembling everything, please post on our Roboard forum as well, because the Roboard engineers keep monitoring the forum and you'll certainly get good info there as well.
Regards
Pedro