4 ms·

Actually, you can do substantially better than one pixel, because the point spread function is bigger than one pixel. Even a technique as simple as centroidin

Actually, you can do substantially better than one pixel, because the point spread function is bigger than one pixel. Even a technique as simple as centroiding will get you sub-pixel accuracy.

sandworm101 8y ago
Yes but to get there you need detectable movement, a change in the image because of motion. A pixel has to change somewhere, the definition of blur. Sub-pixel blur is still blur.
- torrance 8y ago
  GP is right. The point spread function will spread the flux of a point source (eg. a star) across multiple pixels. You can identify the location of the actual star with sub pixel accuracy if you have a good PSF model.
- rcxdude 8y ago
  No, you can get sub-pixel accuracy with a static image. The point is the star takes up multiple pixels, and when it moves by less than a pixel, some pixels get brighter while others get dimmer. This is a common fallacy in computer vision. Often people get to a within a pixel and stop because they assume they can't get better. Most times you can get substantially better. I worked on a qr-code like system where the scanner could reconstruct the code from an image with (slightly) less pixels than 'pixels' in the code.
  - FabHK 8y ago
    Interesting observation. This works presumably because the output of a pixel is not just one bit, but some sort of continuous brightness reading (discretised, of course, sooner or later)? Not obvious a priori, but makes sense.
    - dekhn 8y ago
      https://en.wikipedia.org/wiki/Point_spread_function https://en.wikipedia.org/wiki/Point_spread_function The light is blurred, with a fairly predictable pattern, so if you fit a function to the shape, you can find the peak of the function and that is the most likely center position for the point source.
  - sandworm101 8y ago
    If a pixel gets brighter, that is a moving pixel, ie blur. This is scientific observation. Any change to the data because of movement is blur. The rules for computer vision and astronomy are very different.

Actually, you can do substantially better than one pixel, because the point spread function is bigger than one pixel. Even a technique as simple as centroiding will get you sub-pixel accuracy.

sandworm101 8y ago

Yes but to get there you need detectable movement, a change in the image because of motion. A pixel has to change somewhere, the definition of blur. Sub-pixel blur is still blur.

torrance 8y ago
GP is right. The point spread function will spread the flux of a point source (eg. a star) across multiple pixels. You can identify the location of the actual star with sub pixel accuracy if you have a good PSF model.
rcxdude 8y ago
No, you can get sub-pixel accuracy with a static image. The point is the star takes up multiple pixels, and when it moves by less than a pixel, some pixels get brighter while others get dimmer. This is a common fallacy in computer vision. Often people get to a within a pixel and stop because they assume they can't get better. Most times you can get substantially better. I worked on a qr-code like system where the scanner could reconstruct the code from an image with (slightly) less pixels than 'pixels' in the code.
- FabHK 8y ago
  Interesting observation. This works presumably because the output of a pixel is not just one bit, but some sort of continuous brightness reading (discretised, of course, sooner or later)? Not obvious a priori, but makes sense.
  - dekhn 8y ago
    https://en.wikipedia.org/wiki/Point_spread_function https://en.wikipedia.org/wiki/Point_spread_function The light is blurred, with a fairly predictable pattern, so if you fit a function to the shape, you can find the peak of the function and that is the most likely center position for the point source.
- sandworm101 8y ago
  If a pixel gets brighter, that is a moving pixel, ie blur. This is scientific observation. Any change to the data because of movement is blur. The rules for computer vision and astronomy are very different.