Wiggling Hubble to eek out more resolution

[hjulenissen]

Hubble Ultra-Deep Field - Wikipedia

en.wikipedia.org

https://arxiv.org/pdf/astro-ph/0607632.pdf

I imagine the entire telescope wiggling back and forth to produce slightly sharper images (dont know if that is physically accurate).

Chapter 2.5

"smaller shifts in telescope pointing were applied to different observations at the same position angle.

small-scale dithers were applied to each of the four exposures within a two orbit visit. This dither pattern improved the sampling of the final image by introducing half-pixel off- sets. The ACS/WFC detector critically samples the point spread function (PSF) in the reddest bands but significantly undersamples the PSF at shorter wavelengths. Such undersampling leads to loss of spatial information and aliasing artifacts. The introduction of sub-pixel dithering improves the sampling and allows the reconstruction of a higher-resolution final image and a reduction of artifacts."

Is not the word "dither" unusual in this context? If Hubble could only be pointed in integer-valued angles, I would agree that jumping back and forth between two discrete angles would be "dithering" to produce a (lowpass) dithered fractional position. But this seems to be more like oversampling? Moving the (non ideally sampled) sensor somewhat so as (to be able to) synthesize a more desirable sampling (and to better be able to estimate the information contained within the PSF of the optics)

-h

 

[xpatUSA]

I agree about the "dither" terminology.

As to sampling, I recently compared Lumix DC-G9 80MP mode versus 20MP mode using the slant-edges from DPR's test shots.

In terms of cycles per pixel, the 80MP had much less at MTF9 but, in terms of image-plane lp/mm, it was significantly "better".

 

[JimKasson]

https://en.wikipedia.org/wiki/Hubble_Ultra-Deep_Field

https://arxiv.org/pdf/astro-ph/0607632.pdf

I imagine the entire telescope wiggling back and forth to produce slightly sharper images (dont know if that is physically accurate).

Chapter 2.5

"smaller shifts in telescope pointing were applied to different observations at the same position angle.

small-scale dithers were applied to each of the four exposures within a two orbit visit. This dither pattern improved the sampling of the final image by introducing half-pixel off- sets. The ACS/WFC detector critically samples the point spread function (PSF) in the reddest bands but significantly undersamples the PSF at shorter wavelengths. Such undersampling leads to loss of spatial information and aliasing artifacts. The introduction of sub-pixel dithering improves the sampling and allows the reconstruction of a higher-resolution final image and a reduction of artifacts."

Is not the word "dither" unusual in this context? If Hubble could only be pointed in integer-valued angles, I would agree that jumping back and forth between two discrete angles would be "dithering" to produce a (lowpass) dithered fractional position. But this seems to be more like oversampling? Moving the (non ideally sampled) sensor somewhat so as (to be able to) synthesize a more desirable sampling (and to better be able to estimate the information contained within the PSF of the optics)

Dither is a perfectly good word for this. In a time-based, as opposed to spatially-based system I once designed, I introduced a time-varying voltage offset before the S&H that fed the ADC, and I called that "dither." However, I don't think the lawyers did:

US3999129A - Method and apparatus for error reduction in digital information transmission systems - Google Patents

A method and apparatus for reducing idle channel noise, cross talk and quantizing error noise in a digital information transmission system designed for operation over a predetermined frequency spectrum and having an analog-to-digital converter, an information transmission medium, and a...

patents.google.com

CA1103763A - Signal injection technique - Google Patents

IMPROVED SIGNAL INJECTION TECHNIQUE Abstract of the Disclosure In a digital information transmission system having an analog-to-digital converter (ADC), an information transmission medium, and a digital-to-analog converter (DAC) all located between a transmitting station and a...

patents.google.com