Repository navigation
Astrometry and framing
The Astrometry workspace finds exactly where an image points on the sky (plate solving), labels the objects in it, plans the framing of a target or a mosaic before you image it, and estimates the exposure you need. It has three tabs: Plate solving, Framing & mosaics and Exposure planner.
Plate solving matches the pattern of stars in your image against a star catalogue. The result is a world coordinate system (WCS): a precise mapping between pixel positions and sky coordinates, including the image's centre, scale, rotation and distortion.
- Stars are detected in the image (the brightest 80 are used for matching).
- Catalogue stars around the hint position are projected onto a tangent plane.
- Triangles are formed from neighbouring stars in both sets. The shapes of triangles do not depend on scale, rotation or position, so matching triangles reveal matching stars even when the scale and orientation are unknown.
- The most consistent set of matches gives a first transformation, which is then refined iteratively using all matched stars and fitted as TAN-SIP (a gnomonic projection plus polynomial distortion terms of the chosen order).
- Without a position hint, the solver searches the whole sky, trying the widest field sizes first (60° down to 0.25°) because they need the fewest search positions. Without a scale hint, it tries field sizes from 0.25° to 60°. Mirrored images (taken with a diagonal) are handled.
- Every candidate solution is checked before it is accepted: the pixels must come out square with no skew, the solved field must be about the size that was searched for, and at least 10 stars must match across the whole image. A chance match of a handful of stars fails these checks, so the solver keeps searching and reports "No solution found" rather than a wrong position.
| Solver | Needs | Best for |
|---|---|---|
| Internal + Gaia DR3 (online) | An internet connection the first time for each field (Gaia data are cached), and a rough position | Any field size, including narrow fields of a few arcminutes |
| Internal (local catalog) | The installed star catalogue (HYG, to about magnitude 9, or the built-in bright stars) | Wide fields (a few degrees and up), blind solving, no internet |
| ASTAP (installed) | ASTAP and one of its star databases, installed in the default folder | Fast solving of any field offline |
| astrometry.net (online) | A free API key from nova.astrometry.net in Settings → Observer & accounts | Blind solving of anything, when the other options fail; slower |
When you choose the Gaia solver without a position hint, Sidralune falls back to the local catalogue for the blind search.
Hints make solving faster and more reliable:
| Field | Meaning |
|---|---|
| Hint RA / Hint Dec | Approximate centre. Filled in automatically from FITS headers (RA/DEC, OBJCTRA/OBJCTDEC or an existing WCS) |
| Search radius (°) | How far from the hint to search |
| Scale ("/px, 0=auto) | Image scale in arcseconds per pixel. Filled in from FOCALLEN and XPIXSZ when present. 0 means unknown |
| SIP order | Distortion polynomial order: 0 (none) to 4. 2 suits most telescopes; use 3–4 for camera lenses and wide fields |
| Display | How the image is shown: Auto linked (the default: an automatic stretch that brings out the faint stars of unprocessed data), Auto unlinked (each colour stretched separately) or None (as stored, for images that are already processed). Solving always uses the image's own values |
| Minor body labels | Which minor bodies a solved image labels: All in the field (the default), Bright enough to show, or None. Changing it relabels the image at once, without solving again, and the choice is remembered |
Image scale is 206.265 × pixel size (µm) ÷ focal length (mm). Frames taken by Sidralune's Observatory already contain everything needed.
- Click Open image… (or drop a file onto the viewer). Colour raw images are debayered automatically.
- Check or enter the hints.
- Choose the Solver and click Solve. Stop, available while a solve runs, ends it at once (with ASTAP the ASTAP program is closed too); it also stops an image that is still loading.
The solve runs in the background; the status bar shows progress. While it runs, Open image…, Solve, Write WCS to FITS… and Sync telescope to solution wait for it, so the solution always belongs to the image shown.
With a rough position hint, solving usually takes about a second. A blind solve (no hints) searches the whole sky and can take a few minutes; entering only the scale already makes it several times faster.
The result panel reports the centre, the scale, the rotation angle (and "flipped" for mirror-reversed images), the field size, the number of stars detected and matched, the RMS residual in arcseconds, the time taken (split into finding the stars, solving, placing the labels and measuring the depth; the brightest 1000 stars are used for solving), and the image depth: roughly the magnitude of the faintest stars it shows.
The depth comes from the stars themselves. Catalogue stars are found where the solution puts them and their brightness in the image gives the zero point, which is applied to the faintest 5 % of all the stars detected, so the depth can reach beyond the catalogue's own limit. Every star is measured by fitting the image's own star profile, taken from unsaturated stars: a star whose core is saturated is measured from its wings, so the local catalogue works too, even though its stars are all bright enough to saturate in a deep exposure. Bright enough to show labels the asteroids and comets whose predicted magnitude is at or above that depth. When too few catalogue stars are in the field to tell the depth, all of them are labelled and the result panel says so.
On the image:
- matched stars are circled in green,
- an RA/Dec grid is drawn,
- deep-sky objects (with the PGC galaxies installed, also up to 300 PGC galaxies brighter than magnitude 16 or at least 8 pixels across, the brightest first), named and Bayer stars, and comets and asteroids in the field (positions computed for the image's
DATE-OBS) are marked and labelled, with the minor bodies' predicted magnitudes, - the readout under the image shows RA and Dec at the cursor.
The list of objects in the field is also printed in the result panel, following the Minor body labels choice.
| Button | Action |
|---|---|
| Write WCS to FITS… | Saves a copy of the image with the WCS keywords (CTYPE, CRVAL, CRPIX, CD matrix and SIP coefficients) in its header, readable by any astronomy program |
| Sync telescope to solution | Tells the connected mount where it is really pointing, which corrects its pointing for the next slews |
| Show field in sky chart | Opens the Sky chart in atlas mode, centred on the image, with the image's outline drawn. The outline is traced through the plate solution, so it shows the true rotation and distortion and stays on its patch of sky when you pan or zoom. Showing another solved field replaces it |
The WCS is also stored in the image in memory, so other workspaces (photometry, colour calibration, MPC astrometry) can use it.
- Check the scale: a wrong focal length or pixel size is the most common cause. Set Scale to 0 to let the solver search.
- Increase the Search radius, or clear the hints for a blind solve.
- If a blind solve is slow, enter a rough Hint RA / Hint Dec (even a few degrees off is fine) or the Scale.
- Use the Gaia solver for fields narrower than about 1°; the local catalogue does not have enough faint stars.
- Make sure the image has at least 15–20 well-defined stars: not trailed, not badly out of focus, not dominated by clouds or a bright gradient.
- Try ASTAP or astrometry.net.
The Framing & mosaics tab shows your camera's field of view on a sky chart, so you can compose the image before you take it, and plans mosaics of large targets.
- Type a target in Target and click Update, or use Frame in the Sky chart or the right-click menu's Frame with camera here.
- The rectangle shows the field of the active telescope and camera from Settings → Equipment; Camera / scope shows the field size and image scale.
- Drag the chart to move the frame, and use the mouse wheel to zoom.
- Set Rotation (°) to turn the camera.
- For a mosaic, set Mosaic columns, Mosaic rows and Overlap (%) (15% is a safe minimum for stitching).
The tile list gives the centre of every tile (row, column, RA and Dec). Copy tile list copies it to the clipboard, ready to paste into target containers in the sequencer.
Tiles are laid out on the tangent plane around the mosaic centre and rotated together, so the tiles stay aligned with each other even near the celestial pole. When stacking the tiles, use Mode: Mosaic in Deep-sky stacking.
The Exposure & signal-to-noise planner estimates how long you need to integrate, using the active telescope and camera and your site's sky brightness.
| Field | Meaning |
|---|---|
| Imaging system | The active telescope and camera, with image scale and focal ratio |
| Target surface brightness (mag/″²) | How bright the faint parts you want to show are: about 20–21 for bright nebula cores, 22–23 for galaxy arms and faint nebulae, 24+ for very faint outer structure |
| Sky brightness (mag/″²) | Your site's SQM value (filled in from the active site) |
| Filter bandwidth (nm) | About 88 nm for a luminance or broadband filter per channel, 3–7 nm for narrowband |
| Sub-exposure (s) | Length of each frame |
| Target SNR | Signal-to-noise ratio you want; 10 is a detection, 30 a clean result, 100 very smooth |
The results:
- signal and sky electrons per pixel per sub-exposure,
- the SNR of a single sub,
- the total integration time and number of subs for the target SNR,
- the optimal sub-exposure: the shortest sub for which read noise adds less than 5% to the sky noise; longer subs gain almost nothing but cost more when a frame is lost,
- the critical focus zone in micrometres (how precisely you must focus at your focal ratio),
- the focal length that samples the site's typical seeing at the Nyquist limit for your pixels.
The graph shows the SNR growing with total integration time, with your target SNR marked.
Narrowband filters dramatically reduce the sky background, which is why they work so well from light-polluted sites: compare 88 nm with 7 nm and see how the required time changes.
- Observatory → Imaging → Center and the sequencer's Center target instruction solve and re-slew to put the target in the middle of the frame.
- Observatory → Polar alignment solves three images to measure the polar axis.
- Process → Solve sends an image here; solved images get photometric colour calibration and a WCS grid.
- Photometry → Measure solves frames without a WCS before measuring stars.
- Solar sys. → MPC astrometry measures positions on solved images.
- Scripts can call
PlateSolver.Solvedirectly.
Getting going
Planning and observing
Imaging and processing
Science
- Astrometry and framing
- Photometry and exoplanets
- Spectroscopy
- Solar system
- Radio astronomy
- Databases and citizen science
- Scripting and data analysis
Reference