Python API

Introduction

In order to avoid generating useless files when working with data already in memory, a Python interface is published. The package comes with a library, public interface of which is contained in module api. For example, assuming the photometric channels are NumPy arrays i, y, j h, the RGB (or, rather, BGR image) is rendered as:

from azulero import api as azul

iyjh = np.stack([i, y, j, h])
bgr = azul.process_iyjh(iyjh)

API reference

class azulero.api.Transform(
iyjh_zero_points: tuple = (24.5, 29.8, 30.1, 30.0),
iyjh_scaling: tuple = (2.2, 1.3, 1.2, 1.0),
iyjh_fwhm: tuple = (1.6, 3.5, 3.4, 3.5),
sharpen_strength: float = 0.5,
nir_to_l: float = 0.2,
i_to_b: float = 1.0,
y_to_g: float = 0.5,
j_to_r: float = 0.25,
hue: float = -20.0,
saturation: float = 1.2,
stretch: float = 27.5,
bw: tuple = (28.5, 22.5),
neg_overshoot: float = 0.4,
bgr_curves: tuple = ([(0.5, 0.55)], [], []),
)

Transformation parameters.

Parameters:

  • iyjh_zero_points – Zero points of each channel

  • iyjh_scaling – Scaling of each channel (for white balance)

  • iyjh_fwhm – PSF full width at half-maximum of each channel

  • sharpen_strength – Unsharp masking strength

  • nir_to_l – NIR-to-L rate

  • i_to_b – I-to-B rate

  • y_to_g – Y-to-G rate

  • j_to_r – J-to-R rate

  • hue – Hue rotation angle in degrees

  • saturation – Saturation gain

  • stretch – Stretching parameter

  • neg_overshoot – Negative overshooting parameter

  • bw – Black and white points in AB-mag

  • bgr_curves – Curve adjustment knots for each channel

azulero.api.process_iyjh(
iyjh: ndarray,
wcs: WCS | None = None,
transform: Transform = Transform(),
output: str = '',
) → ndarray

Process an image according to transformation parameters, optionally save intermediate and final images.

Parameters:

  • iyjh – A stack of the I, Y, J, H arrays (e.g. iyjh[1] is the NIR-Y channel).

  • wcs – The WCS parameters or None

  • transform – The transformation parameters

  • output – The output file name (empty string to disable writing)

Returns:

A normalized BGR image (OpenCV layout with bottom-up Y axis).