python3 -m pip install vstarstack
python3 setup.py install [--user]
- sphere mode
- flat mode
In this mode we use knowledge about celestial sphere and the fact, that out images are some projections of celestial sphere. So when we need to perform some movements of image for alignment, we have to use rotations of celestial sphere.
This mode should be used when perspective distorsions are valueable.
- perspective projection - use perspective projection with specified focal length and pixel size
In this mode we don't know anything about celestial sphere and consider images as just flat images. We use standart movements of flat surface (rotation + shift) for alignment.
- stars mode - images contains stars and should be aligned by stars
- compact_objects - images contains some small object, much less than image size, and this object should be cutted out and images should be aligned to center object
Command for work with stars beginning with vstarstack stars.
vstarstack stars detect - detect stars on image
vstarstack stars lonlat
If we use sphere mode, we transform (y,x) coordinates of detected stars into (lat,lon) coordinates, with (0,0) at center of image.
vstarstack stars describe
We build descriptors for N most brightest stars. Each descriptor is invariant to rotations of image - it contains only information about distances to other stars, their relative brightness and angles between pairs of other stars. So we can use this descriptor for identifying stars.
vstarstack stars match
We match the same stars on different images using descriptors
vstarstack stars net
Build net.json - file with info about star matching. This is intermidiate format.
vstarstack stars cluster
Build clusters.json - file with clusters of stars. It contains info about stars coordinates on each frame.
vstarstack stars process - do all steps above in a single run
GNU GPLv3
