get_SNR_map

threadcount.fit.get_SNR_map(cube, signal_idx=None, signal_Angstrom=None, nsigma=5, plot=False)[source]

Create Image of signal to noise ratio in a given bandwidth.

This bandwidth may be selected in 3 different ways:

  1. Choose the indices of the wavlength array to include (signal_idx)

  2. Choose the wavelengths to include (signal_Angstrom)

  3. Have the program fit a gaussian to the data, and choose how many sigmas to include (nsigma). (Uses function: get_SignalBW_idx())

If multiple of signal_idx, signal_Angstrom, and nsigma are given, the order of preference is as follows: signal_idx overrides all others, then signal_Angstrom, and finally the least preferenced is nsigma, which will only be used if either signal_idx or signal_Angstrom are not specified.

Parameters:

  • cube (mpdaf.obj.Cube) – The cube containing data, var, and wave attributes

  • signal_idx (array [int, int], optional) – The indices of the wavelength array to use, by default None

  • signal_Angstrom (array [float, float], optional) – The wavelengths in Angstroms to use, by default None

  • nsigma (float, optional) – Fit a gaussian, and use center wavelength +/- nsigma * sigma, by default 5

  • plot (bool, optional) – Plot the whole image spectrum and highlight the SNR bandwidth, by default False. A tool for troubleshooting/setup.

Returns:

An Image where the pixel values indicate the signal to noise in the selected bandwidth. Given a Spectrum for each spaxel, the SNR for the spaxel is calculated by sum(Spectrum.data)/sqrt(sum(Spectrum.var)).

Return type:

mpdaf.obj.Image

Examples

Given a Cube with name this_cube, then the default bandwidth selection is to fit a gaussian, and use the gaussian center +/- 5*sigma. This is implemented by the following command:

>>> import threadcount as tc
>>> snr_image = tc.fit.get_SNR_map(this_cube)

To use the same method but change the width to, for example, gaussian center +/- 3*sigma, (meaning nsigma=3), then use the following:

>>> snr_image = tc.fit.get_SNR_map(this_cube, nsigma=3)

If you know the specific wavelengths of the bandwidth you would like to use, (for example, 5000-5020 A) then use the following:

>>> snr_image = tc.fit.get_SNR_map(this_cube, signal_Angstrom=[5000,5020])

And finally, if you know the pixel indices (for example, indices 31-60). Note, this is an inclusive range, meaning in this case pixel 60 will be included in the SNR calculation.

>>> snr_image = tc.fit.get_SNR_map(this_cube, signal_idx=[31,60])