The Standard PixInsight Preprocessing Workflow

The preprocessing of astro images, when you get down to it, is the process of calibrating, aligning, and stacking a pile of raw files into one clean base image. There actually aren’t many steps, but what each one does, and why the order is what it is, is worth committing to memory, especially when you’re used to fully automated software: all the more reason to know what it’s actually doing under the hood.

Five Core Steps

PixInsight’s standard preprocessing workflow can be condensed into five steps:

1. Calibration: use calibration frames such as darks, flats, and biases to remove the image’s systematic defects.
2. Cosmetic Correction (removing bad pixels and bad lines): clean up the hot/cold pixels and bad columns that the calibration frames couldn’t eliminate.
3. Debayer (debayering): applies only to color (CFA) cameras, restoring the raw files into a color image.
4. Star Alignment (star registration): align each image to the same reference coordinates.
5. Image Integration: stack the aligned images together to improve the signal-to-noise ratio.

Two points to note about the order: Cosmetic Correction comes after calibration and before Debayer; and Debayer is only done for color cameras, while mono cameras jump straight from calibration and bad-pixel removal to alignment. This workflow is easy to forget over time, so writing it down for reference is the most reliable thing to do.

Why You Should Understand This Workflow

Nowadays many people are used to fully automated software (such as WBPP, or the even more streamlined FBPP) that runs all five steps above in one click. Given that, why bother understanding each step?

Because only by understanding the workflow can you pinpoint which stage broke when preprocessing goes wrong, instead of just shrugging and chalking it up to bad luck. When a stacked image comes out off, someone who understands the workflow can reason backward: was a calibration frame shot badly? Were bad pixels not cleaned thoroughly? Or did the alignment go awry? Someone who doesn’t understand the workflow can only redo the whole thing and hope for the best.

If you want a more intuitive grasp of this workflow, both PixInsight officially and the community have demo videos that walk through the whole thing, from using Blink to inspect the light, dark, flat, and bias frames and cull the obviously broken ones, to calibrating and aligning with WBPP, and then stacking with Image Integration. Watching a video alongside the five steps above makes it clearer what each step actually does on screen.