(Astronomical) Image Registration

Github: Image Registration

TL;DR version

Use chi2_shift to match images.

A toolkit for registering images of astronomical images containing primarily extended flux (e.g., nebulae, radio and millimeter maps). [*]

There are related packages scattered throughout the internet that do the same thing, but with different features.

The general goal is to align images that look kind of like these:

The input image The input image shifted and corrupted with gaussian noise

Quick Example

Given two images, shift the second to match the first and check that they are the same. This notebook also shows examples.

from image_registration import chi2_shift
from image_registration.fft_tools import shift
xoff, yoff, exoff, eyoff = chi2_shift(image, offset_image, noise,
...                                   return_error=True, upsample_factor='auto')
corrected_image2 = shift.shiftnd(offset_image, -yoff, -xoff)

Module APIs:

image_registration Package

fft_tools Package (and a description of the Fourier Image Manipulation Tools)

The most successful of the methods implemented here is chi2_shift

There is an ipython notebook demonstration of the code here and in pdf here

Indices and tables

[*]Apparently astronomical images look a lot like microscopic images. So maybe this code is good for coaligning bacteria!