NSLS-II-CSX / csxtools

The calculate_flatfield function has an issue when it is used with the current state of the CSX detector.

The problem is that more than 1/2 the detector is insensitive to light, so the median is not representative of the true data. The line below then means that the data and limits get re-scaled to unreal values.

csxtools/csxtools/utils.py

Line 268 in 52ff964

limits = np.nanmedian(image) * limits

Thanks @mpmdean . We are working on code for one big PR for various issues. For others that run into this prroblem, you may use the work around below by setting half=True.

import logging
logger = logging.getLogger(__name__)
import numpy as np

from csxtools.image import rotate90, stackmean
from csxtools.utils import  calculate_flatfield, get_images_to_3D,  get_fastccd_images

def get_fastccd_flatfield(light, dark, flat=None, limits=(0.6, 1.4), half=False, half_args = (7, 486)):
    """MODIFIED from csxtools original: Calculate a flatfield 
    This routine calculates the flatfield using the
    :func:calculate_flatfield() function after obtaining the images from
    the headers.
    Parameters
    ----------
    light : databroker header
        The header containing the light images
    dark : databroker header
        The header from the run containin the dark images
    flat : flatfield image (optional)
        The array to be used for the initial flatfield
    half : calculate for just the "good" half - hard coded
        Default is False
    half_args : Tuple for exluding entire sides of detector (left versus right)
        Left side is refers the the left side of the image after raw data is 
        processed with get_fastccd_images().  Default arguments are for the left side
        using the FrameStore mode.
    Returns
    -------
    array_like
        Flatfield correction
    """
    images = get_images_to_3D(get_fastccd_images(light, dark, flat))
    images = stackmean(images)
    if half == True:
        #rows because "super columns" are the large 10 pixel bins, but camera is on side.
        row_start, row_stop = half_args
        images[:,row_start:row_stop] = np.nan
        #plt.figure()
        #im = plt.imshow(images, vmin =0, vmax = 500)
        #cbar = plt.colorbar(im)
        #cbar.set_label('ADU gain corrected')
    flat = calculate_flatfield(images, limits)
    removed = np.sum(np.isnan(flat))
    if removed != 0:
        logger.warning("Flatfield correction removed %d pixels (%.2f %%)" %
                       (removed, removed * 100 / flat.size))
    return flat

Issue #77 covers this

I didn't spot that prior version of the same issue. I think this can be closed.

flat field issue