Question about downsampling.

Started May 3, 2012 | Discussions thread
crames
Regular MemberPosts: 188
Like?
Re: Thanks all!
In reply to John Sheehy, May 6, 2012

John Sheehy wrote:

Personally, I don't accept any level of resolution that would change the detail with any change of registration. By my standard, real resolution requires 6 pixels per line pair, or 3 pixels per line width. Anything much less is distortion; not resolution.

In sampling theory, there is no restriction on registration or phase. As long as the Nyquist criterion is met the image at the focal plane can be reconstructed exactly. You're saying that Nyquist was off by a factor of 3?

Note the necessity of reconstruction for sampling to work as expected, without distortion. A 2x up-sample is usually good enough.

http://oai.dtic.mil/oai/oai?verb=getRecord&metadataPrefix=html&identifier=ADA310889

http://www-cs.ccny.cuny.edu/~wolberg/pub/crc04.pdf

Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow