Stars in the sky

[ka9q]

I think there are a couple of genuine stars in there, but I think most of the small dots may be damaged pixels from cosmic ray hits. Some guys from Johnson told me last year that they periodically have to discard and replace cameras because of accumulated radiation damage, so this might be it.

I think the astronauts could compensate for these to a limited extent by regularly taking dark- and flat-field images to determine where the bad pixels are, but apparently they don't do that.

I wonder what this does to the sensors in Hubble. It's in a lower inclination and slightly higher orbit  that would accumulate less radiation, but it still accumulates some.

[Chew]

The pic has no metadata. There is also a star on the dark side of the Moon.

[cjameshuff]

I wonder what this does to the sensors in Hubble. It's in a lower inclination and slightly higher orbit  that would accumulate less radiation, but it still accumulates some.

It's also got radiation-hardened sensors that may be better shielded, and are probably subject to frequent recalibration.

[raven]

Ok, I could be wrong, and you all make some good points.

[Trebor]

I could be just displaying my ignorance here, but is the black curve along the bottom of the aurora not the nightside horizon of the Earth? And if it is, why are there "stars" visible between the camera and what should be the Earth?  All those white dots look more like some kind of artifact to me.

Or am I just looking at it wrong?

The white dots are artefacts, as pointed out bad pixels through radiation damage on the camera sensor is the most likely cause.
If you ever see a raw unprocessed Hubble image (or any spacecraft) they also tend to show the same thing.

Like in this Cassini image for example :
http://saturn.jpl.nasa.gov/multimedia/images/raw/casJPGFullS47/N00128399.jpg

[ka9q]

That certainly explains it.

What's the usual practice for compensating for these things? How often are dark and flat-field test frames taken?

[Zakalwe]

That certainly explains it.

What's the usual practice for compensating for these things? How often are dark and flat-field test frames taken?

Certainly for any decent amateur astro-photography the answer is "all the time". Flats are required to remove vignetting and "dust-bunnies" from images. Darks are required to remove sensor noise and hot pixels. In addition, we also use bias (offsets) to remove read-out noise from the sensors. To help control the sensor noise we cool the imaging chip using peltier coolers. My camera runs at about -30 C for this very purpose.

Darks can be taken and then "banked". This is one of the many attractions of using a dedicated cooled CCD camera in comparison to an off-the-shelf DSLR. As you can control the temperature of the sensor, you can create a bank of darks which can be used over and over (sensor noise varies with temperature). This is much more difficult to do with a DSLR.

Flat fields have to be created per image (indeed, per filter if doing RGB imaging), as the orientation of the camera and the dust spots will vary from image to image.


https://blogs.stsci.edu/newsletter/2011/10/20/wfc3-flat-fields/

[ka9q]

Interesting, so Hubble can be pointed at the sunlit earth, at least with certain filters.

[Trebor]

Interesting, so Hubble can be pointed at the sunlit earth, at least with certain filters.

Here's one of them....


And here is a Hubble mosaic image of the sun :
http://nicmosis.as.arizona.edu:8000/UVFLOOD/WFPC_SUN_100.jpg

[ka9q]

The SUN??!

[Tanalia]

It's all done with mirrors.

[ka9q]

Ah. That makes sense. It also explains the irregular changes in brightness over the solar disc.