Re: Spot size in relation to pixel size
- From: AES <siegman@xxxxxxxxxxxx>
- Date: Wed, 30 May 2007 21:34:53 -0700
In article <1180573969.347264.17030@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>,
wadexkelman@xxxxxxxxxxx wrote:
The relationship between the spot size and the pixel size depends upon
what you are trying to accomplish, but in general, if you want to
maximize your resolution, put two pixels under the spot. That is
based upon the Nyquist sampling theorem.
You can make a hand-waving argument to see this. If your lens has
a spot size that is huge compared to your pixels, you won't have as
much resolution across your chip as the chip can resolve. If your
pixels are huge compared to your spot size, you won't have as much
resolution as the lens can produce. If we took a guess at the middle
range and set the spot size equal to the pixel size, we might assume
that the resolution is at a maximum. And it would be, if our subject
were a picket fence that is just resolved, with the white boards one
spot size wide and occupying one column of pixels and the black spaces
occupying the adjacent pixels. That would be great, until the camera
was pointed 1/2 pixel to one side, at which point the entire fence
would go gray and there'd be no resolution whatsoever. To overcome
this, put two pixels under the spot.
A *great* explanation (truly).
[And as a side question: Isn't there a fair amount of dead real estate
between pixels, on some chips anyway?]
.
- Follow-Ups:
- Re: Spot size in relation to pixel size
- From: Dave Martindale
- Re: Spot size in relation to pixel size
- References:
- Spot size in relation to pixel size
- From: jfld22
- Re: Spot size in relation to pixel size
- From: wadexkelman
- Spot size in relation to pixel size
- Prev by Date: Re: Spot size in relation to pixel size
- Next by Date: Re: Fiber Core Imaging
- Previous by thread: Re: Spot size in relation to pixel size
- Next by thread: Re: Spot size in relation to pixel size
- Index(es):
Relevant Pages
|
|
