Subject:
|
Re: Color images? (was: Re: digital camera entirely made of mindstorms?)
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 28 Sep 2000 15:53:18 GMT
|
Viewed:
|
1424 times
|
| |
|
|
I tried this a little when I first made my robot using the red bat wings that
come with the set, and I couldn't get a good image through one of them. One of
the big problems is that the sensor is mostly infrared, with some red
sensitivity. I haven't measured it, but I don't think it can see green or blue
very well. Using real photo filters would work better than Lego parts, but the
image distortion wouldn't be too much of a problem, since you could place the
colored part directly over the light sensor.
You could print out filters on transparancy stock with an inkjet and try it.
The colors of Cyan Magenta and Yellow would let more light through than Red
Green Blue. CMY filters will produce a very noisy color signal, but it might
be detectable.
In lugnet.robotics, David C. Pyatt writes:
> In theory, yes I believe it could be done, however I would be concerned that a
> 1x2 filter would be so thick as to distort the image drastically... better to
> take a sheet of transparency material and print a red, green and blue spot and
> attach it possibly to a wheel/pulley (sorry I'm at work and not in full Lego
> mode) the 6 holed variety and rotate it with a rotation sensor to lock in the
> correct positioning.... ?
>
> I had an article on how to due color scanning with a BW hand scanner that was
> a multipass using filters... I think rotating the filters in and having the
> exact lens positioning would work much better.
>
> DaveP
>
> In lugnet.robotics, David Eaton writes:
> > In lugnet.robotics, David Eaton writes:
> > > Ok, maybe I'm just out in left field on this one, or maybe someone will
> > > actually be inspired to try it out... How about making a color digital • camera?
> > >
> > > Proposal: modify the existing design for the camera by adding on a colored
> > > filter, and an extra motor. Bascially, create a wheel at the end of the lens
> > > which had trans-red, trans-green, and trans-blue pieces. For each pixel that
> > > the camera takes, cycle the filter for R, G, & B values. You'd have to tweak
> > > the results a lot, but could it be done?
> > >
> > > DaveE
> >
> > Ok, so it's uncouth to reply to my own messages, but more on the color
> > blending:
> >
> > -each filter could be 2 1x2 trans bricks-- the red filter would be green &
> > blue, green would be red & blue, blue would be red & green
> > -you'd also have a 4th filter (a hole) for a baseline gamma value. Hence, you
> > could determine how much additional gamma you'd need to add for weighting the
> > filtered values. So if you got an RGB of 10,30,20 (all out of 100, not 255) • and
> > a gamma of 80 (out of 100), you could bump up the RGB to something like
> > 20,40,30 (well, ok, you'd have to figure out the percentages, but that's just
> > for example)
> >
> > DaveE
|
|
Message has 2 Replies:
Message is in Reply To:
16 Messages in This Thread:
  
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
Active threads in Robotics
|
|
|
|
|
|
