Why Is My Stacked Image Green?

You stack a night’s frames, open the result, and the whole thing is green. Not slightly green, but a flat, swampy, uniform green with a few white specks in it. It looks as if the stacking failed.

It did not. That green is what a color camera records, and every stacked image from every one-shot-color sensor starts out looking like that. Here is why, and what happens to it.

The sensor has twice as many green pixels

A color camera does not record color. It records brightness through a mosaic of tiny filters (red, green and blue) laid over the pixels in a repeating pattern. In almost every sensor that pattern is two green pixels for every red and every blue. The arrangement is named after its inventor, Bryce Bayer, and it copies the human eye, which is most sensitive to green.

So before anything else happens, half of your data is green. Then two more things push the same way. The green filter passes the widest band of wavelengths, so a green pixel collects more light per second than a red or blue one behind the same lens. And the sky itself, light pollution and airglow, is strongest in the green and yellow part of the spectrum.

Add those up and the green channel in a raw stack is routinely one and a half to two times brighter than red or blue. Displayed as-is, that is a green image.

Your camera normally hides this

A daylight photograph from the same camera is not green, because the camera applies a white balance before you ever see it: it multiplies the red and blue channels up until a gray card looks gray. That correction is baked into every JPEG and applied in every raw converter by default. You have never seen your camera’s uncorrected output, which is why the stack looks so alarming.

Astrophotography software works from the uncorrected data on purpose. The daylight white balance is wrong for a night sky. It was calibrated for sunlight reflected off things, not for starlight and airglow, so the correction has to be derived from the data itself.

What “removing the green” means

There are two separate corrections, and it matters that they are separate.

Neutralizing the background shifts each channel so the empty sky is the same brightness in all three. This is a subtraction, not a multiplication. It moves the sky to gray and leaves everything brighter than the sky exactly as it was. The green cast of the sky goes away, and a green nebula would stay green.

White balancing multiplies the channels so that a known-neutral thing is neutral. The problem is finding the known-neutral thing. Daylight photography uses a gray card. Astrophotography uses stars: either the typical star in the field, on the argument that most stars are roughly sun-colored, or, better, specific stars whose colors are in a catalog. Akastroid does the first by default and the second when your frames record where the telescope was pointing. (Why the choice of reference star matters.)

Only after both does the image have color that means anything.

Why Akastroid shows you the green

The split view in the Processing tab puts the raw stack on one side and the result on the other, and the raw side is green. That is deliberate, and an earlier version of the app quietly neutralized the raw side because the green looked like a mistake.

It was put back because the green is the real data. Removing it properly means measuring what the sky and the stars are and correcting from that, rather than pushing a “less green” slider. A before-and-after that starts from a pre-cleaned image shows almost no difference and hides that step.

If the green side bothers you, drag the divider. It is not a fault. A raw stack from a color camera that is not green would be more surprising.

The short version

• A color sensor has twice as many green pixels, the green filter passes the most light, and the sky glows green. A raw stack is green.
• Your camera applies a daylight white balance to hide this. Astrophotography works from the uncorrected data because the daylight balance is wrong for the sky.
• The fix is two steps: shift the sky to gray, then scale the channels so a reference star is neutral.
• The green “before” image in Akastroid is the real starting point, not a bug.