Align a sequence of astronomical FITS files based on sources extracted in each image. The aligned files will be geometrically reprojected so all the images are the same size and shape.
It is designed to work exclusively with reduced data from Las Cumbres Observatory. FITS data from LCO is Rice compressed and contains sources extracted using SEP during our data pipeline processing in a catalogue (CAT
) FITS Header Data Unit (HDU).
Download this repository :
python setup.py install
from fits_align.ident import make_transforms
from fits_align.align import affineremap
from glob import glob
from numpy import shape
tmp_dir = "<FULL PATH TO INPUT FILES>"
img_list = sorted(glob(os.path.join(tmp_dir,"*.fz")))
ref_image = img_list[0]
images_to_align = img_list[1:]
identifications = make_transforms(ref_image, images_to_align)
If you have FITS files with image data you could use the following to reproject them.
aligned_images = [ref_image]
for id in identifications:
if id.ok:
alignedimg = affineremap(id.ukn.filepath, id.trans, outdir=tmpdir)
aligned_images.append(alignedimg)
If you just have FITS files which contain on photometry catalogues, which have pixel coordinate values (e.g. x, y) that you want to align (ie. no image data):
import pandas as pd
identifications = make_transforms(img_list[0], img_list[1:], hdu='CAT')
catalogues = []
for i, catfile in enumerate(img_list):
with fits.open(catfile) as hdul:
data = pd.DataFrame.from_records(hdul['CAT'].data)
if i != 0:
(matrix, offset) = identifications[i-1].trans.matrixform()
newcoords = np.dot(data[['x','y']], matrix) + offset
(x,y) = np.transpose(newcoords)
data.update({'x':x, 'y':y})
catalogues.append(data)
If you are using a none standard place for your catalogue HDU, that can be passed in a parameter to the make_transforms()
function i.e.
identifications = make_transforms(ref_image, images_to_align, hdu=0)
This is a customised fork of AliPy by Malte Tewes. I have removed the dependency on SciPy in favour of pure NumPy (for linear algebra) and Pillow (for image array transforms).