To: Mimas{at}aol.com, atm{at}shore.net, tkrajci{at}san.osd.mil
From: tkrajci{at}san.osd.mil
Reply-To: tkrajci{at}san.osd.mil


> From:  Don Ware - Mimas{at}aol.com
> On 1/12/03, Tom Krajci wrote:
>
> >Recently there has been some discussion of robo Foucalt to test
mirrors,
> >test reflectivity,etc.
> >
> >Has linearity of the camera been validated?
>
> As the camera is only checking for relative intensity between the
average of
> pixels from the left zone and pixels from the right zone, is linearity
> important?

Are you sure the camera/optical system/test rig/light source has no
vignetting?  How have you checked for this?  If there is vignetting or
other uneven illumination of the setup...you may *think* you've got equal
relative intensity between different locations/zones...but....  :-(

> I have been using RoboFoucault extensively and do see changes in
intensity
> from zone to zone, even in a spheroid, but this change is equal on both
sides
> so I still get a null at what I feel is the correct position.

Interesting.  I'd be wary if you didn't get even illumination from a sphere
on the test stand.  At least you say you get even/equal values on both
side...but to what degree of precision have you measured this?  You may
*feel* you have a correct setup, but I'd rather verify it by objective
means...unless you imply that your feelings are truly subjective in this
matter.  (Mine probably aren't.  That's why I don't trust them.)

As I used to sign my email on this list when I was a Captain....
"Trust!...but verify!"

Along related lines...is there any published analysis of
RoboFoucalt...especially in terms of sensitivity analysis of video cameras.
 Are video cameras up to the task of detecting the lighting differences to
an accurate enough degree so that this test setup can measure precisely and
reliably enough?

Tom Krajci
Tashkent, Uzbekistan

--- BBBS/NT v4.00 MP