AstroBin is the home of astrophotography - Share full resolution images, preserve acquisition details, discover equipment, discuss techniques, and learn from the community.Discover AstroBin

Community tools and Hobby projects

Astronomy tools
Tony GondolaAlain ESCAFFREchvvkumar
•64 replies•3.3k views
Christian Brendlin avatar
Christian Brendlin avatar

Hello,

I have developed a lightweight tool called ophi for analyzing and sorting FITS light frames before stacking.

The basic idea is that you point it at the folder containing your light frames, and it measures things such as star count, FWHM, HFD, eccentricity, sharpness/blurriness, SNR, clipping, background uniformity/variance, cloud/fog indicators, trailing, and other possible artifacts. It then generates JSON and HTML reports, and can sort the frames into accept/review/reject folders based on the results.

It works well on my own dataset, but that is also the limitation: I only have my own camera/filter setup to test against.

I would mostly be interested in hearing from a small number of people willing to try it on real-world data and let me know:

  • number of frames processed

  • camera / image size

  • filters used

  • processing time

  • whether the accept/review/reject decisions seemed reasonable

  • obviously misclassified frames

  • crashes or FITS files it could not read

It is completely free, and I am not asking anyone to change their workflow. I am simply trying to find out where it works well and where it fails.

As a rough benchmark, on my system it analyzed and reported on 348 ASI2600MM light frames in about one minute using NVMe storage. I would be very interested to see how it performs on other machines and other datasets.

Thank you to anyone willing to test it. Criticism is much more useful than compliments, especially examples where it gets things wrong.

If a few people would like to try it, I can share the current build and setup instructions.

Well written Helpful Respectful Concise Engaging
Tony Gondola avatar
Tony Gondola avatar

Christian Brendlin · May 21, 2026, 01:32 PM

Hello,

I have developed a lightweight tool called ophi for analyzing and sorting FITS light frames before stacking.

The basic idea is that you point it at the folder containing your light frames, and it measures things such as star count, FWHM, HFD, eccentricity, sharpness/blurriness, SNR, clipping, background uniformity/variance, cloud/fog indicators, trailing, and other possible artifacts. It then generates JSON and HTML reports, and can sort the frames into accept/review/reject folders based on the results.

It works well on my own dataset, but that is also the limitation: I only have my own camera/filter setup to test against.

I would mostly be interested in hearing from a small number of people willing to try it on real-world data and let me know:

  • number of frames processed

  • camera / image size

  • filters used

  • processing time

  • whether the accept/review/reject decisions seemed reasonable

  • obviously misclassified frames

  • crashes or FITS files it could not read

It is completely free, and I am not asking anyone to change their workflow. I am simply trying to find out where it works well and where it fails.

As a rough benchmark, on my system it analyzed and reported on 348 ASI2600MM light frames in about one minute using NVMe storage. I would be very interested to see how it performs on other machines and other datasets.

Thank you to anyone willing to test it. Criticism is much more useful than compliments, especially examples where it gets things wrong.

If a few people would like to try it, I can share the current build and setup instructions.

Please do, I’m always culling data so I’d be interested.

Frédéric Ruciak avatar
Frédéric Ruciak avatar

Hi Santiago I'll try it for 😃.CS Frédéric

Christian Brendlin avatar
Christian Brendlin avatar

Tony Gondola · May 21, 2026, 02:16 PM

Christian Brendlin · May 21, 2026, 01:32 PM

Hello,

I have developed a lightweight tool called ophi for analyzing and sorting FITS light frames before stacking.

The basic idea is that you point it at the folder containing your light frames, and it measures things such as star count, FWHM, HFD, eccentricity, sharpness/blurriness, SNR, clipping, background uniformity/variance, cloud/fog indicators, trailing, and other possible artifacts. It then generates JSON and HTML reports, and can sort the frames into accept/review/reject folders based on the results.

It works well on my own dataset, but that is also the limitation: I only have my own camera/filter setup to test against.

I would mostly be interested in hearing from a small number of people willing to try it on real-world data and let me know:

  • number of frames processed

  • camera / image size

  • filters used

  • processing time

  • whether the accept/review/reject decisions seemed reasonable

  • obviously misclassified frames

  • crashes or FITS files it could not read

It is completely free, and I am not asking anyone to change their workflow. I am simply trying to find out where it works well and where it fails.

As a rough benchmark, on my system it analyzed and reported on 348 ASI2600MM light frames in about one minute using NVMe storage. I would be very interested to see how it performs on other machines and other datasets.

Thank you to anyone willing to test it. Criticism is much more useful than compliments, especially examples where it gets things wrong.

If a few people would like to try it, I can share the current build and setup instructions.

Please do, I’m always culling data so I’d be interested.

Hey Tony! I have messaged you via DM.

Gerald avatar
Gerald avatar

Hi everyone,

I want to present “Rejector”, a tool to review FITS/XISF files and to reject based on metrics and/or on manual override. Rejector is a lightweight Windows tool, focused on performance.

What it does

  • Loads FITS / XISF frames (mono & OSC) from a folder, tested also with 150 Megapixel dataset (IMX411)

  • Detection and rotation of the image orientation

  • Computes per-frame metrics: FWHM, eccentricity, star count, background, satellite-trail confidence, and an overall quality score

  • Filter-aware: handles mixed L / R / G / B / Ha / OIII / SII datasets in one session

  • PixInsight-style STF stretching

  • ROI loupe with Ctrl+drag square selection at any zoom

    Free & open source: https://github.com/photon1503/blink-o-mat

    📷 image.pngimage.pngVideo Project 6.gif

Well written Helpful Respectful Concise Engaging
Vin avatar
Vin avatar

astroscore — AI-powered craft feedback for astrophotography images

A tool I've been building that uses Claude AI to analyse astrophotography images and return structured, parameter-by-parameter feedback calibrated against AstroBin's IOTD, Top Pick, Top Pick Nomination, and general image population.

What it does: Returns scores across 8 craft parameters (signal-to-noise, star quality, processing, composition, colour palette, detail & resolution, gradient & background, noise character), a composite weighted score, and specific feedback (based on its data analyses) on what craft factors would push an image toward TP or IOTD level.

What it doesn't do: Judge art. Your image is your image — the tool is a structured rubric, not an arbiter of aesthetics. Disagree with it freely (I do, and I built it!) because everyone's aesthetic is different and valid.

Optional inputs: object name, processing workflow, imaging setup (backyard/remote), feedback language (12 languages supported), and image stage (finished or WIP).

Free to use: astroscore.leelaastroimaging.org

Running costs are real (API and cloud infrastructure) — there's a buy me a coffee button if you find it useful, entirely optional.

Happy to answer questions here.

Well written Helpful Concise Engaging
Minhong Suh avatar
Minhong Suh avatar

Hello guys!

I created an android astrophoto stacking app named astrostacker about 3 months ago. It is still in development but still it has some cool features for a mobile app like Plate solving or Narrowband stacking. The app is currently on google playstore for free with no ads at all. It is a bit big with a size around 1.3gigabytes since it has all the necessary star catalouges and assets needed for the app to work offline. I would appreciate all feedback for the app and would try my best to catch your needs. Thanks for reading and have a good day :)

Helpful Respectful Concise Supportive
MaksPower avatar
MaksPower avatar

ASIAir has something quite similar with a filter that uses altitude and azimuth.

Well written
Joerg avatar
Joerg avatar

Free ISS TIANGON HUBBLE Transit over Moon / Sun - pushover Infoservice based on own location

Hi all,

over the past months I've been working on an open-source tool for predicting aircraft transits across the Sun and Moon (using local ADS-B data — that part only works as a local install, since aircraft trajectories are only reliable a few minutes ahead).

However, while building it, some folks asked me and I added satellite transit predictions (ISS, Tiangong, Hubble) almost as an afterthought — and ended up using that part far more than I expected. Unlike aircraft, satellite orbits are stable enough to predict transits a day or two in advance, so I set it up to send myself a push notification whenever a transit crosses my home imaging location. No more manually checking transit-finder sites every few days and still missing events etc.

Since it was already running for me, I made it available as a simple website pushover service for anyone who wants to use it. (Free, no account, no registration)

You just enter your Pushover app key (Pushover is a generic push notification app if you don't already have it yet) and your coordinates, and you'll get an alert 1–2 days before ISS, Tiangong or Hubble crosses the Sun or Moon as seen from your location. Predictions are recalculated every 30 minutes against fresh TLEs, so they track orbit adjustments.

If you'd like to use it here the link: Pushover Info Service Satellite transit alerts — ISS · Hubble · Tiangong

The whole thing (including the aircraft transit part) is open source on GitHub, in case anyone wants to look under the hood or run it themselves.

However, I'd genuinely appreciate feedback — especially on prediction accuracy from locations far from mine (I'm in northern Germany), since that's hard for me to verify on my own. And if there are other satellites worth adding, I'm open to suggestions.

Clear skies, Joerg

📷 transit-infoservice.jpgtransit-infoservice.jpg

Well written Helpful Respectful Engaging
Jeffrey Horne avatar
Jeffrey Horne avatar

I built a small free Mac helper app for ASIAIR users that ramps ZWO camera cooling more gradually instead of jumping straight to the final setpoint. The goal is to help users in humid conditions test whether slower cooldown reduces camera-window/sensor condensation.

It’s a hobby project, not commercial. I’m looking for a few ASIAIR + cooled ZWO camera users, especially in humid areas, who have seen dew or condensation during cooldown and would be willing to test and report results.

DM me for more info…must be macOS / apple silicon

Well written Respectful Concise
samuel krieger avatar
samuel krieger avatar

Hi everyone,

I'm the developer of AstroBurst, a free and open-source (GPLv3) desktop app for processing astronomical images. I've been building it solo for a while, and this is the first time I'm putting it in front of a community that will really put it through its paces, so your feedback would mean a lot.

It runs fully offline on Windows, macOS, and Linux. No account, no cloud, no subscription, no locked features.

The workflow it covers, end to end:

  • Calibration: bias / dark / flat with master frames, automatic dark scaling by exposure ratio

  • Debayer: OSC frames (RGGB/BGGR/GRBG/GBRG, auto-detected from headers), bilinear or super-pixel

  • Subframe selection: gates frames by FWHM, eccentricity, SNR, and star count, then weights the survivors for the integration

  • Stacking: sigma-clipped integration with auto-alignment (FFT phase correlation or star-based affine + RANSAC)

  • Drizzle: per-channel, with Square / Gaussian / Lanczos3 kernels

  • Background extraction: polynomial gradient and light-pollution removal (subtract or divide)

  • Color calibration: SPCC against real Gaia DR3 (via VizieR)

  • Stretching: GHS (Generalized Hyperbolic), arcsinh, and a star-protected masked stretch, plus linked or per-channel STF

  • Enhancement: Richardson-Lucy deconvolution (with empirical PSF estimation), à-trous wavelet denoise, SCNR

  • Narrowband: filter auto-detection and palette mapping (SHO / HOO / Foraxx / Hubble) resolved by wavelength

  • Analysis: plate solving (astrometry.net), interactive aperture photometry with optional Gaia cross-match, star detection, FFT, deep-zoom tiles

A few things that might make it worth a look:

  • All pixel data stays in 32-bit float through the whole pipeline, and the composite is non-destructive (white balance and SCNR always reconstruct from the originals).

  • The heavy lifting is in Rust and the preview renders on the GPU (WebGPU), so STF adjustments stay real-time even on large frames.

  • It reads FITS (including multi-extension) and ASDF. As far as I know it's the first non-Python tool to read Roman Space Telescope ASDF files, so reprocessing public JWST/Hubble/Roman data goes through the same pipeline as your own captures.

  • Exports PNG (8/16-bit) and FITS with WCS and metadata preserved, plus PROGRAM/HISTORY provenance cards.

Honest status: it's at v0.5.x. Usable end to end, but young, and there are surely rough edges I haven't hit yet. Bug reports, "here is how my workflow actually goes," and feature requests are exactly what I'm hoping for.

Repo, downloads, and full-pipeline screenshots: https://github.com/samuelkriegerbonini-dev/AstroBurst

Thanks for taking a look. Clear skies.HST Pillars of Creation narrowband composite previewAnalysis panelHeader explorer with channel assignment

Well written Helpful Respectful Engaging
Frank Adler avatar
Frank Adler avatar

Hi everyone,

I have published my first PixInsight script: ParallaxStarLayers.

The script takes a processed astrophotography image split into a starless image and a stars-only image, plus a linear version of the image for astrometric data. It then uses GAIA catalogue data available in PixInsight to identify stars, retrieve distance information where possible, and sort the stars into configurable distance buckets.

For each bucket, the script exports a separate PNG star layer. It also exports the starless base image and an additional “unknown” layer for stars that could not be matched or do not have usable distance information.

The resulting layers can be used for video animations, subtle parallax effects, or interactive image presentations where the star field is separated by approximate distance rather than treated as one flat layer.

I use this, for example, for the hero images on my astrophotography detail pages:

https://adfr.io/astro/20260522_wr134

The script requires GAIA data inside PixInsight. At minimum, GAIA DR3-SP is needed, but the full GAIA DR3 dataset gives better results because it contains significantly more stars. The dataset can be downloaded via the PixInsight Software Distribution and configured under:

Processes → GAIA → Settings

The script can be installed through my PixInsight update repository:

https://pixinsight.adfr.io/

Please note that the trailing slash is important.

To install it:

  1. Open Resources → Updates → Manage Repositories

  2. Add https://pixinsight.adfr.io/

  3. Run Resources → Updates → Check for Updates

  4. Install the available package

A note about the current state: the script is not signed yet, so PixInsight will show a warning during installation. I am aware of this and currently looking into the signing process while learning what is required to become a certified PixInsight developer.

This is an early release and there is certainly room for improvement, but it already works well enough for my own workflow. Feedback, suggestions, and bug reports are very welcome.

I also wrote a short post about the script here:

https://adfr.io/thoughts/20260624_parallaxstarlayers_pixinsight_script

Well written Helpful Respectful
Robert Gillette avatar
Robert Gillette avatar

Looks fascinating! Eager to give it a try.

CS, Bob