I want to be able to detect whether the wearer of some prototyp-glasses is looking on some specific point. In the stage now it can be triggered by something easy like lets say the color RED.
So how could i realize that?
i just dont know how good this sensor would detect some color in a certain distance which would be crucial for detecting whether the glasses are pointed at the point.
Also interesting are the video-shields like this one:
As i understood they kind of have low resolution, right? Couldnt find the resolution of this specific one. Also i cant find a way to get it here in europe, since i am from germany.
But i am also not sure whether for me the small resolution really is a problem. It would be enough if he could detect my point somehow - whether over a color code or somehow else.
I think that would work with such an video-shield, or what do you think?
But maybe you guys do find a much easier way to make that work or have some help to my ideas above.
Thank you already in advance!
But i am also not sure whether for me the small resolution really is a problem.
Resolution probably isn't going to be your problem with the VideoExperimenter - lack of any colour information (or, indeed, most luminance information) most likely is.
Oh i didnt see that.
Is there actually a Video-Shield or sth like that, that would provide any colour-Informations?
Or maybe somebody has an idea how to make that work completly different.
My task here is just to have the information, whether the wearer of the glasses is heading towards an object or not.
It would be perfect if there would be a way to do so, without having to plug the glasses to a computer.
Color sensing may be easier done with an Android phone, using the camera. There are examples in the Android SDK and possibly programs that you can use.
If you are going the Arduino approach, maybe you can use a gyro / acceleration combo to determine the position and orientation of the glasses and calculate the direction the person is looking at from that, rather than colour? There are some fairly cheap IMU boards you can find, along with tutorials on using them.
Both of those are for short-range viewing and have poor resolution, and
aren't going to be adequate to "detect some color in a certain distance".
For that, you probably will need some sort of video camera and processing
over the X-Y pixel array.
Maybe the video experimenter shield might work with a monochrome camera with an optical filter over it, but you'd need a fairly large area emitter at longer ranges to compensate for the limited capture resolution.