3 ms·

> The authors say that the 570 megapixel Dark Energy Camera (DECam) may be useful for follow-up observations. I was curious what kind of resolution you'd have

> The authors say that the 570 megapixel Dark Energy Camera (DECam) may be useful for follow-up observations. I was curious what kind of resolution you'd have at this distance but not sure I did the math right. The camera has a resolution of 0.27"/pixel[1] which is 0.000075 degrees. Then to get size at 500AU -> tan((pi/180) * 0.000075)(500 149597870700) ~98megameters, which is like 8 earth diameters. Is this right? [1]: https://www.darkenergysurvey.org/the-des-project/instrument/the-camera/ https://www.darkenergysurvey.org/the-des-project/instrument/...

georgeburdell 1y ago
Sub pixel resolution is possible by observing the space over time
- DoingIsLearning 1y ago
  In astronomy how is image registration done? Is there some sort of astronomic dead reckoning system? Or is it just image based with some kind of Homography technique?
  - ddahlen 1y ago
    Most images are registered by finding the location of known stars from a high resolution catalog such as GAIA. So you fit centroids of all stars in an image, then do a search of Gaia sources in that location and do some sort of linear or higher order polynomial mapping of pixel space to the celestial sphere reference coordinates ICRF. Astrometry.net is a service that does an approximate version of this on the web. I am glossing over many details here, but this is roughly what happens.
    - DoingIsLearning 1y ago
      Very cool to learn about other areas of knowledge. Astrometry.net really made this quite visual thanks. For the curious wanting to go down this rabbit hole here seems to be an ESA archive dataset of 'GAIA Data Release 3' (released 3 June 2022) : https://cdn.gea.esac.esa.int/Gaia/gdr3/ https://cdn.gea.esac.esa.int/Gaia/gdr3/
perihelions 1y ago
I think that's about right, but the real purpose isn't resolving spatial details on the planet (which you correctly found isn't possible), but to find the planet at all—to distinguish a point of light out of background noise. A fine angular resolution is still helpful, for SNR reasons: smaller pixels contain less noise, wheras the signal pixel contains the same amount of signal!
ddahlen 1y ago
Basically all observations of solar system objects besides the big planets are going to be unresolved point sources. IE: We don't have telescopes which can resolve most objects into an image. An example, Hubble famously had to image a ton to get a grainy, ~12 pixel approximation of Pluto. We can do a lot with measurements of points, even reconstruct 3d models if we have enough data. Resolving power is related to the PSF (Point Spread Function) size. The PSF is the image on your detector if you have an infinitely small point source. A quick google search says that DECam has a PSF of at least an arcsecond (atmosphere is probably causing issues for that). Which means anything smaller than an arcsecond is going to be unresolvable. However, you still want pixels smaller than your PSF, since PSFs are typically gaussian-ish, having a bunch of measurements within the gaussian allows you to estimate the center accurately. This is vital for Astrometry (the measurement of position).

georgeburdell 1y ago

Sub pixel resolution is possible by observing the space over time

DoingIsLearning 1y ago
In astronomy how is image registration done? Is there some sort of astronomic dead reckoning system? Or is it just image based with some kind of Homography technique?
- ddahlen 1y ago
  Most images are registered by finding the location of known stars from a high resolution catalog such as GAIA. So you fit centroids of all stars in an image, then do a search of Gaia sources in that location and do some sort of linear or higher order polynomial mapping of pixel space to the celestial sphere reference coordinates ICRF. Astrometry.net is a service that does an approximate version of this on the web. I am glossing over many details here, but this is roughly what happens.
  - DoingIsLearning 1y ago
    Very cool to learn about other areas of knowledge. Astrometry.net really made this quite visual thanks. For the curious wanting to go down this rabbit hole here seems to be an ESA archive dataset of 'GAIA Data Release 3' (released 3 June 2022) : https://cdn.gea.esac.esa.int/Gaia/gdr3/ https://cdn.gea.esac.esa.int/Gaia/gdr3/

perihelions 1y ago

I think that's about right, but the real purpose isn't resolving spatial details on the planet (which you correctly found isn't possible), but to find the planet at all—to distinguish a point of light out of background noise. A fine angular resolution is still helpful, for SNR reasons: smaller pixels contain less noise, wheras the signal pixel contains the same amount of signal!

ddahlen 1y ago

Basically all observations of solar system objects besides the big planets are going to be unresolved point sources. IE: We don't have telescopes which can resolve most objects into an image. An example, Hubble famously had to image a ton to get a grainy, ~12 pixel approximation of Pluto. We can do a lot with measurements of points, even reconstruct 3d models if we have enough data. Resolving power is related to the PSF (Point Spread Function) size. The PSF is the image on your detector if you have an infinitely small point source. A quick google search says that DECam has a PSF of at least an arcsecond (atmosphere is probably causing issues for that). Which means anything smaller than an arcsecond is going to be unresolvable. However, you still want pixels smaller than your PSF, since PSFs are typically gaussian-ish, having a bunch of measurements within the gaussian allows you to estimate the center accurately. This is vital for Astrometry (the measurement of position).