Usage

Experimental data

Retrieve data on xarray format with the functions from cmod_functions. Then load the data for example:

import xarray as xr

path_to_data = ... # Change this to your path
ds = xr.open_dataset(path_to_data)

Then estimate_velocity_field() will return a class containing estimated velocity data of all pixels. Different estimation parameters can be set with a EstimationOptions class, which is accepted as an arguments of estimate_velocity_field(). By default a cross-correlation technique is used to estimate time delays, but this can be changed by setting EstimationOptions.method.

eo = tdve.EstimationOptions()
print(eo.method)

This will print TDEMethod.CC, indicating that a cross-correlation method (CC) is being employed. In addition to the own parameters in the EstimationOptions class, further time delay estimation options are specified in EstimationOptions.cc_options, EstimationOptions.cca_options and EstimationOptions.ccf_options. If you use cross-correlation time-delay estimation method (by setting method=TDEMethod.CC), you only need to specify EstimationOptions.cc_options, and likewise for the other methods.

print(eo.cc_options)

With output CC Window: None, Minimum CC Value: 0.5, Running Mean: True, Running Mean Window Max: 7, Interpolate: False.

Check out the documentation of the option class for the time delay estimation method you plan to use to get an overview of the parameters. All parameters have default values, but most likely you will want to set some of these by your own.

Some important options are:

  • eo.use_3point_method If False, use 2 point method to estimate velocities from time delays.

  • eo.cc_options.cc_window Size of the window of time lags considered for the cross-correlation function.

  • eo.cc_options.interpolate Should interpolate ccf to find the maximum (if set to False, the time delays are restricted to integer multiples of the sampling time).

Estimate velocities by:

movie_data = tdve.estimate_velocity_field(u.CModImagingDataInterface(ds), eo)
vx = movie_data.get_vx()
vy = movie_data.get_vy()
confidences = movie_data.get_confidences()
R = movie_data.get_R()
Z = movie_data.get_Z()

These contain the estimated velocities components for each pixel together with the confidence of the estimation (a measure of uncertainty). The velocity field can be plotted with standard plotting libraries:

axes_size = cosmoplots.set_rcparams_dynamo(plt.rcParams, num_cols=1, ls="thin")
plt.rcParams["mathtext.fontset"] = "custom"

fig = plt.figure()
ax = fig.add_axes(axes_size)
norm = mpl.colors.Normalize(vmin=0, vmax=1)
qiv = ax.quiver(
    R,
    Z,
    vx,
    vy,
    confidences,
    scale=210000,   # Scale arrows
    scale_units="xy",
    angles="xy",
    norm=norm,
)
cbar = fig.colorbar(qiv, format="%.2f")
cbar.ax.set_ylabel(r"$CC_{max} $")
ax.set_xlabel("R [cm]")
ax.set_ylabel("Z [cm]")
ax.set_ylim(np.min(Z) - 0.5, np.max(Z) + 0.5)
ax.set_xlim([np.min(R) - 0.5, np.max(R) + 0.5])
plt.show()
_images/velocity_field.png