add_noise_in_pixels

lstchain.image.modifier.add_noise_in_pixels(rng, image, extra_noise_in_dim_pixels, extra_bias_in_dim_pixels, transition_charge, extra_noise_in_bright_pixels)

Addition of Poissonian noise to the pixels

Parameters:

rngnumpy.random.default_rng

Random number generator

image: `np.ndarray`

Charges (p.e.) in the camera

extra_noise_in_dim_pixels: `float`

Mean additional number of p.e. to be added (Poisson noise) to pixels with charge below transition_charge. To be tuned by comparing the starting MC and data

extra_bias_in_dim_pixels: `float`

Mean bias (w.r.t. original charge) of the new charge in pixels. Should be 0 for non-peak-search pulse integrators. To be tuned by comparing the starting MC and data

transition_charge: `float`

Border between “dim” and “bright” pixels. To be tuned by comparing the starting MC and data

extra_noise_in_bright_pixels: `float`

Mean additional number of p.e. to be added (Poisson noise) to pixels with charge above transition_charge. This is unbiased, i.e. Poisson noise is introduced, and its average subtracted, so that the mean charge in bright pixels remains unaltered. This is because we assume that above transition_charge the integration window is determined by the Cherenkov light, and would not be modified by the additional NSB noise (presumably small compared to the C-light). To be tuned by comparing the starting MC and data

Returns:

image: np.ndarray

Modified (noisier) image