Space has fascinated me since before I knew what infrastructure was. There's something about pointing optics at a smudge of light 1,300 light-years away and pulling an image out of it that connects the engineer in me to something much older. This post is about how I do that — from a backyard in a light-polluted suburb, on a budget that doesn't require selling a server rack.
The Gear
The entry tax for astrophotography sounds steep until you realize how much the software can compensate for hardware limitations. My current setup:
- Telescope: Celestron StarSense Explorer LT 114AZ — a 114mm reflector, perfectly adequate for nebulae and galaxies
- Camera: ZWO ASI120MC-S — a color planetary/deep-sky camera under $150
- Mount: The stock alt-az mount, manually tracked (no goto, no motor — more on this later)
- Filter: A cheap dual-band narrowband filter, which is the single best purchase for light-polluted skies
- Computer: A Raspberry Pi 4 running KStars/EKOS for capture control
None of this is exotic. The dual-band filter is the force multiplier — it cuts sodium and mercury light pollution while passing H-alpha and OIII wavelengths, where most nebula emission lives.
Acquisition
I capture with KStars + EKOS on the Raspberry Pi. It's open-source, runs the INDI device framework, and interfaces with the ASI camera natively. For a manual mount, the workflow is simple: polar align as best you can, center the target, and fire a sequence of short exposures.
Without tracking, you're limited to exposures short enough that the stars don't trail. With a 114mm scope at f/8, that's roughly 15-20 seconds before stars elongate noticeably. Stacking compensates for what tracking would otherwise give you. — Field notes, November 2025
For the Orion Nebula session, I captured 120 frames at 15 seconds each, plus calibration frames: 30 darks (same exposure, lens cap on), 30 flats (uniform white light source), and 30 bias frames (minimum exposure, cap on). These calibration frames are how you remove sensor noise, amp glow, dust shadows, and optical vignetting from your final image.
Stacking with Siril
Siril is free, open-source, and genuinely excellent. The workflow for a standard deep-sky integration:
- Load your light frames, darks, flats, and biases into their respective directories
- Run the preprocessing script — Siril will calibrate, register (align stars across frames), and stack automatically
- The output is a 32-bit FITS file with the combined signal of all your exposures
Registration is where Siril earns its reputation. It identifies stars across frames and aligns them sub-pixel accurately, so slight tracking drift between frames disappears in the stack. 120 mediocre frames become one excellent one.
Processing
Post-stacking in Siril involves background extraction (removing the light pollution gradient), color calibration against known star catalogs, and stretching — taking the linear FITS data and expanding the dynamic range so the nebula becomes visible.
For final output I export to TIFF and do a light pass in GIMP — mostly contrast adjustments and a mild sharpen. The goal is to represent the object accurately, not to manufacture something Instagram-ready.
Results
The Orion Nebula from a Bortle 7 sky, 30 minutes of total integration, dual-band filter, manual mount. The Trapezium cluster at the core resolved cleanly. The nebulosity structure — the wings, the dark lane — was unmistakably there. On a screen in a warm house, rendered from photons that left that gas cloud when the Roman Empire was a going concern.
The cost of the whole setup was under $500. The marginal cost of each imaging session is zero. If you've ever looked up and wondered if you could reach back — you can. The tools are there, and most of them are free.