Turning discarded astrophotography data into discoveries with SpacePixels

As astrophotographers, we spend so much time using rejection algorithms during stacking to perfectly erase anything that moves across our frames. But I started thinking about all the dynamic data we are throwing away, like faint asteroids, satellites, comets, streaks and other slow moving objects that happen to drift through our target fields. It turns out those so called problems are also signals. And once you start looking for them, they have a way of looking back - enter, "SpacePixels".

Turning discarded astrophotography data into discoveries with SpacePixels

SpacePixels is a dedicated tool to explore that lost data. It takes your aligned, monochrome FITS sequences and runs them through a custom detection engine to hunt down moving targets. If you keep color data, you are covered since it supports importing color and mono sixteen bit FITS images. This is not a telescope control suite or a processing behemoth. It is a lean investigator, the friend who shows up after you stacked the pretty picture and says, mind if I see what hit the cutting room floor.

What it actually does under the hood

The first job is separating the living from the still. Stars do not go anywhere across a short sequence, and hot pixels and column defects are stuck like barnacles. SpacePixels uses a custom track linking algorithm that learns the local sky and links discrete detections into coherent motion paths, while filtering out stationary stars and the usual sensor gremlins. In practice, that means it can surface faint movers in the mess of faint stars without flooding you with false hits.

You do not have to spend your evening twiddling thresholds either. Smart Auto Tuning lets you feed the app a handful of frames and then it does the math. It evaluates seeing, star elongation, and small guiding jitters, and proposes sensible extraction settings for that session. You can still take the wheel, but you start from a spot that respects your sky rather than a generic preset.

Before detection, the tool gives you a small workbench. You can batch convert color FITS to mono when you want a uniform luminance stream. You can batch stretch the sequence for a clean preview, then blink through with simple arrow key navigation. A lot of would be finds reveal themselves when you can page through a calm, consistent stretch and see a speck step across the field.

Solving is built in as well. SpacePixels integrates with ASTAP for offline solving and with Astrometry.net for online solving when you want to annotate frames or confirm where a track sits on the sky. The plan is to use solves later for automatic annotation of detections so your report does not just show a streak but names its stage.

Once the pipeline finishes, you get a clean HTML report and a set of animated, cropped GIFs for every candidate. The reports include object centric and star centric views, which is a fancy way of saying you can stabilize on the object to confirm its motion or on the stars to confirm the object is not an artifact that dances with the noise. That visual confirmation step is where science meets the human eye, and it is honestly the best part.

How this changes the nightly workflow

The nice thing about this approach is that it bends around your current habits rather than demanding a new religion. Keep shooting deep integrations on your favorite targets. Keep stacking for the wall print. Then pass the aligned frames to SpacePixels and ask a different question. Not how do I make the nebula pop, but what moved through my field while I was staring elsewhere.

In my case, I tried a sequence of a galaxy shot at a modest focal length. The stack was fine. The rejects looked ordinary. The report, though, surfaced a faint point that walked diagonally across a hundred frames, barely above the noise floor in any single exposure. I had never seen it in manual blinks, because I was not trained to look for motion at that speed and that signal to noise. But the linked track popped in the object centered animation, and the coordinates checked out. That is a small discovery to be sure, but it changes the way I score a night. I no longer measure success only by the final image. I also count the stories my data almost threw away.

Moving from pretty pictures to productive pictures

For a long time, the amateur community has leaned into aesthetics, and rightly so. Beauty is what gets us outside on cold nights. But beauty and discovery are not rivals. A well run session can do both. If you are a backyard imager, a citizen scientist, or simply a curious soul with a tracker and a few clear hours, you can turn your normal routine into a small survey. The barrier to entry is low when the software respects your existing files and your patience.

SpacePixels is free and it is offered in the spirit of exploration. It is still evolving, and you can already download a beta to experiment with. Think of it as an extra sense for your data rather than a finished product. Tools grow best when people use them and talk plainly about what works and what does not. If you try it and share your results, the whole community gets a little sharper.

Practical notes for the careful tinkerer

You will get the most out of the tool if you feed it well aligned sequences with consistent exposure and calibration. Flats and darks still matter because cleaner data means fewer false flags. If you are mixing multiple nights, sort by session so the auto tuner can read each run on its own terms. Expect to do a bit of triage on the first passes. The reports make that painless, and the animations quickly train your eye to spot grains of truth from patterns in the noise.

Plate solving is worth enabling once you have a feel for the pipeline. Knowing exactly where a track crossed your field invites follow up, and it makes your notes more useful to others. If you find something that looks fresh, you can compare coordinates and motion with existing catalogs or share the findings for independent checks. None of that requires a ladder of credentials. It only requires care.

Why this matters

We live in a sky that is busy in ways our final images rarely reveal. Every cleared satellite trail, every rejected streak, and every faint moving dot is a sign that the field is alive. That realization does not reduce the romance of a dim nebula or a dusty galaxy. It adds another kind of wonder, a practical one. A stack is not only a path to a framed print. It is also a record of the night, a ledger of what passed in front of you while you tried to steady the mount and sip lukewarm coffee. SpacePixels helps you read that ledger. When you do, you might find that even the throwaways are worth keeping.

Where can I find it?

The project and source can be found at https://github.com/ppissias/SpacePixels