Astro
DEENFR

Learn · Technique

Lucky Imaging — sharp through shimmering air

September 1, 2026 · 8 min read

Lunar and planetary photography is deep-sky with the signs reversed. Light is abundant — the Moon is so bright that exposures of hundredths of a second are enough. The adversary is the air: kilometers of shimmering atmosphere smear every detail long before the optics reach their limit. The answer to that is called lucky imaging — photographing like a goalkeeper: not looking longer, but more often, and keeping only the best moments.

Freezing the air

Seeing — the atmospheric unrest that makes stars twinkle — changes on a timescale of hundredths of a second. A long exposure averages all these distortions into a soft mush. A video with short exposure times, by contrast, freezes each individual state of the atmosphere: most frames are bent, squashed, smeared — but again and again one of them happens to catch a calm moment in which the telescope gets to show what it can do. The software then rates thousands of frames, throws ninety percent away and stacks the rest — selection beats patience.

Three steps of lucky imaging. A long exposure averages the atmospheric unrest and smears every crater. A video of hundreds of short frames freezes the turbulence: most frames are distorted, a few catch a calm moment. Only these best frames are selected and stacked — the result is sharper than any single image.60 s exposurethe average of the turbulence:every crater turns to mush→the lucky hitvideo: 1/100 s per framethe turbulence is frozen —some moments happen to be calm→the best 10% stackedsharp AND low-noise —selection beats patience
The name says it all: you do not wait for good seeing, you fish the lucky moments out of bad seeing. Stacking the best frames then also pushes down the noise — the same square-root statistics as in deep-sky work, only with images from a single minute.

Jitter as a tool: drizzle

While stacking, something remarkable happens along the way: no frame hits the pixel grid in exactly the same place — seeing and tracking shift the image constantly by fractions of a pixel. What sounds like a nuisance is information: the drizzle algorithm, originally developed for the Hubble Space Telescope, computes with these sub-pixel shifts and reconstructs the image on a finer grid than the sensor actually provides.

Drizzle: two coarsely sampled shots of the same detail, shifted against each other by half a pixel between frames, together carry more information than either one alone. The drizzle algorithm distributes them onto a grid twice as fine and reconstructs detail smaller than a camera pixel.frame 17+frame 18 — offset by ½ pixel→reconstructed on a twice as fine gridDrizzle — the jitter becomes a toolBecause every frame hits the pixel grid slightly differently, hundreds of framestogether hold detail below the pixel size — originally developed for Hubble
Hundreds of minimally shifted coarse grids together yield a fine one: the plight of jitter turns into resolution. It only works because videos deliver frames in abundance — one more reason why the Moon is filmed rather than photographed.

A few things worth knowing about this:

  • The Barlow lens is the scale control. It extends the focal length by a fixed factor — here a 2.5× Barlow turned the f/4 Newtonian into an f/10 system. The goal is the right sampling: the smallest detail the optics deliver should fall on about two pixels. As a rule of thumb, aim for an f-number of roughly five times the pixel size in micrometers — beyond that, more focal length only yields darker images, not sharper ones.
  • Planetary cameras are specialists for speed. Small sensors with small pixels, read out at dozens to hundreds of frames per second — with exposures of hundredths of a second, cooling is not needed. The 2012 lunar panoramas in this gallery were made with a simple full-HD planetary camera whose small field was guided across the Moon tile by tile.
  • The mosaic is the price of the scale. If you image the Moon that large, each video covers only a section. The finished panorama is a patchwork of individually stacked tiles — each one a complete lucky-imaging result in its own right, joined together in Photoshop at the end.
  • The aperture sets the hard limit. Lucky imaging removes the seeing limit, but not diffraction: an 8-inch telescope resolves about 0.6 arcseconds at best — roughly one kilometer at the Moon. The Apollo descent stages remain fundamentally hidden from any Earth-bound telescope, no matter how good the night is.
  • That is why at the Moon the air counts, not the light. The best lunar night is not the clearest but the calmest — a veil of haze hardly matters, a jittering jet stream ruins everything. What seeing forecasts can do is described in the night article.
  • And what the Moon shows is decided by the light. Terminator or full Moon, relief or albedo — that is not a question of technique but of illumination. It belongs to the Moon article.

In our own sky

Both lunar entries in this gallery are pure lucky imaging, built tile by tile from videos:

Waning three-quarter Moon — relief in grazing light78 objects — show
The three-quarter Moon panorama from 2012: GSO 8″ Newtonian, 2.5× Barlow, full-HD planetary camera — many videos, from each of them only the sharpest moments, assembled into a mosaic of 4116 × 6174 pixels.My photograph · Moon · view gallery entry
Plato — the dark crater with the polished floor3 objects — show
The field test of the method: Plato with its smooth lava floor. The tiny craterlets on it have for generations been the classic touchstone for optics and seeing — that they are hinted at here is the achievement of the ninety percent that were discarded.My photograph · Moon · view gallery entry
Copernicus — the spoked wheel above Mare Insularum9 objects — show
Copernicus in almost full light: no shadow relief left, but rays and the finest albedo structure — details at the limit of what aperture and atmosphere jointly allow.My photograph · Moon · view gallery entry

Further reading

← All articles