Subject:
|
RE: Color images? (was: Re: digital camera entirely made of mindstorms?)
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 28 Sep 2000 20:12:04 GMT
|
Viewed:
|
822 times
|
| |
|
|
***Warning*** Non-Lego part discussion to follow.....
Are we looking for a light sensor that can distinguish color, or really
just a color sensor (that by human definition works in the visible
spectrum)? Hamamatsu makes a variety of small (6mm x 8mm) single color
photodiodes for color detection.
Spectral sensitivity:
Blue 400nm to 540nm
Green 480nm to 600nm
Red 590nm to 720nm
All three are nominally insensitive to IR, response time 0.5us.
Now I haven't tried this myself, but.....I could imagine making a "lens"
out of three of the fiber optic elements, sheathing them in some of the
darker Lego tubing, and running the other end to the face of the
photodiodes inside some suitable dark brick (like maybe a battery box). I
haven't thought far enough ahead to figure out the interfacing. You'd
probably need a bit of signal conditioning....
Jeff
Jeffrey Hazen
North Mill Technology
jeffrey.hazen@northmill.net
-----Original Message-----
From: David C. Pyatt [SMTP:dcpyatt@netzero.net]
Sent: Thursday, September 28, 2000 3:21 PM
To: lugnet.robotics@lugnet.com
Subject: Re: Color images? (was: Re: digital camera entirely made of
mindstorms?)
..and you are going to share this great sensor, aren't you? :)
Dave
In lugnet.robotics, John Barnes writes:
> I have been experimenting, on and off, with a "better light sensor"
> for the RCX. No offense meant to Lego or any of the more purist
> Legoists, but I indulge myself in designing various other sensors
> for RCX use.
>
> In the course of that indulgence, I have been working on a color sensor.
> It is conceptually similar to the Lego light sensor in that it supplies
> its own light, but in my case it has four LEDs, a red, green, blue and IR.
> It has a standard silicon sensor (phototransistor) and it energizes each
> LED in turn as well as sampling with no light. It can therefore subtract
> the background level from each of the four levels from the LEDs. It uses
> the return from the IR to decide if an object is present and then measures
> the signal from each color to decide upon a color value to return to the
> RCX. 0 - 9 is no object in range, 10 - 89 encodes the spectrum and
> 90+ is returned for white. It all seemed quite easy to start off with • except
> for the one thing that seems to be dogging this thread;
>
> The typical phototransistor seems to have quite a poor response from
> green though the blue. In my case, the response to green and blue
> is about one tenth what it is at red and about 1/25 what it is at IR. It
> varies from one device type to another however.
>
> The other observation is that the colors produced by the average color • printer
> don't behave the same as those embedded in materials like Lego bricks.
> I have a spectrum printed on strip of paper wrapped around a large Lego
> tire which I rotate to use as a continuously varying color source for
> testing. Its
> green and lego brick green which look the same do not have similar • properties
> when illuminated by my LED system. So if you are planning on trying to use • a
> color printer to make color filters for your camera, you may not get quite
> the results
> you're expecting ... or is it the LED light ? .....
>
> JB
>
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
Active threads in Robotics
|
|
|
|
|
|
