Normalizing Radial Gradient Filters

This example showcases the filters found in the “radial” module.

These are:

• Normalizing Radial Gradient Filter (NRGF) (sunkit_image.radial.nrgf)

• Fourier Normalizing Radial Gradient Filter (FNRGF) (sunkit_image.radial.fnrgf)

• Radial Histogram Equalizing Filter (RHEF) (sunkit_image.radial.rhef)

import matplotlib.pyplot as plt

import astropy.units as u

import sunpy.data.sample
import sunpy.map

import sunkit_image.radial as radial
from sunkit_image.utils import equally_spaced_bins

sunpy sample data contain a number of suitable FITS files for this purpose. Here we will use a sample AIA 171 image.

aia_map = sunpy.map.Map(sunpy.data.sample.AIA_171_IMAGE)

/home/docs/checkouts/readthedocs.org/user_builds/sunkit-image/conda/stable/lib/python3.12/site-packages/astropy/io/fits/hdu/image.py:634: VerifyWarning: Invalid 'BLANK' keyword in header.  The 'BLANK' keyword is only applicable to integer data, and will be ignored in this HDU.
  warnings.warn(msg, VerifyWarning)

Both the NRGF and FNRGF work on radial segments above their application radius.

Here we create those segments radial segments. Each segment created will be of equal dimensions radially. The distance between 1 solar radii and 2 solar radii is divided into equal parts by the following two lines.

radial_bin_edges = equally_spaced_bins(1, 2, aia_map.data.shape[0] // 4)
radial_bin_edges *= u.R_sun

base_nrgf = radial.nrgf(aia_map, radial_bin_edges=radial_bin_edges, application_radius=1 * u.R_sun)

We will need to work out a few parameters for the FNRGF.

Order is the number of Fourier coefficients to be used in the approximation. The attenuation coefficients are calculated to be linearly decreasing, you should choose them according to your requirements. These can be changed by tweaking the following keywords: mean_attenuation_range and std_attenuation_range.

order = 20
base_fnrgf = radial.fnrgf(
    aia_map,
    radial_bin_edges=radial_bin_edges,
    order=order,
    application_radius=1 * u.R_sun
)

Now we will also use the final filter, RHEF.

base_rhef = radial.rhef(aia_map, radial_bin_edges=radial_bin_edges, application_radius=1 * u.R_sun)

/home/docs/checkouts/readthedocs.org/user_builds/sunkit-image/conda/stable/lib/python3.12/site-packages/sunkit_image/utils/utils.py:359: RuntimeWarning: Mean of empty slice
  mid = np.nanmean(in_array)

Finally we will plot the filtered maps with the original to demonstrate the effect of each.

fig = plt.figure(figsize=(15, 15))

ax = fig.add_subplot(221, projection=aia_map)
aia_map.plot(axes=ax, clip_interval=(1, 99.99) * u.percent)

ax1 = fig.add_subplot(222, projection=base_nrgf)
base_nrgf.plot(axes=ax1)
ax1.set_title("NRGF")

ax2 = fig.add_subplot(223, projection=base_fnrgf)
base_fnrgf.plot(axes=ax2, clip_interval=(1, 99.99) * u.percent)
ax2.set_title("FNRGF")

ax3 = fig.add_subplot(224, projection=base_rhef)
base_rhef.plot(axes=ax3)
ax3.set_title("RHEF")

ax.coords[0].set_ticklabel_visible(False)
ax1.coords[0].set_ticklabel_visible(False)
ax1.coords[1].set_ticklabel_visible(False)
ax3.coords[1].set_ticklabel_visible(False)

fig.tight_layout()

plt.show()