Astrometry improvements
Astrometric enhancements
List of stuff planned to be looked into in the coming months regarding astrometry. Each group of items will probably be a MR to avoid changing too many files at same time. (Point 0 will still require important changes).
- Done in !579 (merged)
- Refactor catalogue management
- Pass most online queries through TAP to avoid parsing + add new catalogues (exoplanets #1171 (closed) etc)
- Add all downloaded catalogues to a cache folder and improve cache searching to void too many queries.
- Review nomad command / rename / merge with solsys
- Add proper motion in the display (and in the catalogues in general)
Objective: Have a consistent process for both local catalogues, annotation catalogues and downloaded catalogues
- Done in !591 (merged)
- Reconcile wcsdata and wcslib (remove wcsdata)
- Update libwcs (add build in Windows CI)
Objective: Avoid redundant data between the 2 structs that forces keeping them synced. Avoid having to introduce distorsion terms in wcsdata. Now this structure holds only ra/dec and if image has been platesolved.
- Done in !610 (merged)
- Fix conventions (+0.5)
- Fix solving at pole
In Siril internal solver, use same bounds for scale as in asnet (instead of hard-coded +/- 0.2)
Objective: Use consistent conventions with other softwares (DS9). Have a robust solution close the poles
- Done in !612 (merged)
- Add SIP in header write-up
- Add SIP in Siril internal solver
- Make sure we loop to project finally at center (internal solver)
- Start with linear solution then increase the order
Objective: Have astrometric solutions valid accross the whole image. Handle write-up of the new keys in the headers
- Done in !631 (merged)
-
Change match_catalog so that we don’t start with a projected list (will make it easier to loop)done in !579 (merged) -
Avoid temp files tmp/catalog.projdone in !579 (merged) - when plate solving with a local catalog, search for a radius as large as given in the settings for local asnet if not found at first, perhaps with a spiral search
Objective: Continue refactoring the astrometric process and find a solution when the target is off
- & 6. Done in !636 (merged)
- Full blind solve from asnet (at least scale, most probably jpg too)
- Refresh header after platesolving (avoids having to save)
- Avoid saving new sequence for seqplatesolve
(needs to address #944) - Refactor astrometry GUI to allow sequence platesolving as well
Note: #944 will need to be addressed more generically to save one hdu on a fitseq. Could be useful for other features like updating FITS keywords as well.
Objective: Finish refactoring the astrometric process and find a solution when the target is unknown
- Done in !643 (merged)
- Add astrometric alignment (first step of mosaics, just clean current branch)
- Done in !643 (merged)
- From SIP data, compute inverse distorsion (checkout lensfun to use same convention? Second step of mosaics?)
Objective: Add astrometric alignement as a new registration method and undistort images with inverse SIP coefficients
- Add computed cometary registration
Longer term
- Online request to asnet (requires API key setting in prefs)
- create a Gaia DR3 extract for local use
- Read ASTAP catalogues?
Need triage:
- UI elements for conesearch/show, make a dialog
-
Prefs for annotations colorsdone in !590 (merged) - Add a catalogue cache strategy
- Separate online query code (fetch_url etc) to another file. Make them common to the check for siril updates
- Still need to gather the idles of the different commands to a dedicated file and avoid gtk in the algo files
- As part of the conesearch UI it would be nice to be able to set a preference in the settings window for the default observatory (implemented in !669 (merged))
- Need to query other catalogues if SIMBAD returns wrong photometry. Example with Kepler-76
- Add disto to UI