Speaking as someone who ran the software development for automated inspection of ordnance at SAIC:
You always want to characterize your sensors to the greatest degree that resources permit. In this case, where you've just spent billions to get the sensor into service, and you are dealing with scientific data, characterization of marginal pixels is a minor expense compared to the benefit. On Sun, Aug 26, 2012 at 1:27 PM, Abd ul-Rahman Lomax <[email protected]>wrote: > At 11:10 PM 8/23/2012, [email protected] wrote: > >> In reply to Terry Blanton's message of Wed, 22 Aug 2012 21:15:15 -0400: >> Hi, >> [snip] >> >Dead moving pixies? >> > >> >http://www.dailymail.co.uk/**sciencetech/article-2192008/**Are-Martian-<http://www.dailymail.co.uk/sciencetech/article-2192008/Are-Martian->overlords---just-dead-pixels- >> **camera-Images-beamed-**Curiosity-lead-talk-UFOs-Mars.**html >> > >> >T >> ..surely if they were dead pixels they would be at the same position >> relative to >> the frame in each image? >> >> Clearly that is not only not the case, but pixels there were "dead" in >> one image >> are suddenly "live" again in the next???? >> > > Yes. Pixels in image sensors can be marginal or noisy, plus we may be > seeing artifacts of how images are combined from pixels. Analysis of the > raw data would show, through statistical analysis. > > The "frame," because of vibration, may not be eactly the same with each > frame capture. > > How bright are stars on Mars? A star might raise the level of a pixel > closer to flipping the bit. It might never be able to flip two pixels, if > the image is sharp. > > Trying to figure this out from jpeg files can be ridiculous, because of > compression and loss of primary pixel data. One would want the raw data. > And it's quite possible that image compression has taken place on Mars, so > that the necessary data transmission is reduced. > > I would guess that Mission Control is capable of testing this all. The > positions of bright stars would be known, for example. I guess they could > arrange transmission of some block of primary image data. > > If it's worth it! Is it? >
