def test_margin():
size = (10, 10)
edge = 1
mask1 = mask.margin(size, edge)
mask2 = np.zeros(size)
mask2[:, :edge] = 1
mask2[:, -edge:] = 1
mask2[:edge, :] = 1
mask2[-edge:, :] = 1
mask2 = mask2.astype(bool)
assert_array_equal(mask1, ~mask2)
def mask_img(img, geo,
edge=30,
lower_thresh=0.0,
upper_thresh=None,
bs_width=13, tri_offset=13, v_asym=0,
alpha=2.5,
tmsk=None):
"""
Mask an image based off of various methods
Parameters
----------
img: np.ndarray
The image to be masked
geo: pyFAI.geometry.Geometry
The pyFAI description of the detector orientation or any
subclass of pyFAI.geometry.Geometry class
edge: int, optional
The number of edge pixels to mask. Defaults to 30. If None, no edge
mask is applied
lower_thresh: float, optional
Pixels with values less than or equal to this threshold will be masked.
Defaults to 0.0. If None, no lower threshold mask is applied
upper_thresh: float, optional
Pixels with values greater than or equal to this threshold will be
masked.
Defaults to None. If None, no upper threshold mask is applied.
bs_width: int, optional
The width of the beamstop in pixels. Defaults to 13.
If None, no beamstop polygon mask is applied.
tri_offset: int, optional
The triangular pixel offset to create a pointed beamstop polygon mask.
Defaults to 13. If None, no beamstop polygon mask is applied.
v_asym: int, optional
The vertical asymmetry of the polygon beamstop mask. Defaults to 0.
If None, no beamstop polygon mask is applied.
alpha: float or tuple or, 1darray, optional
Then number of acceptable standard deviations, if tuple then we use
a linear distribution of alphas from alpha[0] to alpha[1], if array
then we just use that as the distribution of alphas. Defaults to 2.5.
If None, no outlier masking applied.
tmsk: np.ndarray, optional
The starting mask to be compounded on. Defaults to None. If None mask
generated from scratch.
Returns
-------
tmsk: np.ndarray
The mask as a boolean array. True pixels are good pixels, False pixels
are masked out.
"""
if tmsk is None:
working_mask = np.ones(img.shape).astype(bool)
else:
working_mask = tmsk.copy()
if edge:
working_mask *= margin(img.shape, edge)
if lower_thresh:
working_mask *= (img >= lower_thresh).astype(bool)
if upper_thresh:
working_mask *= (img <= upper_thresh).astype(bool)
if all([a is not None for a in [bs_width, tri_offset, v_asym]]):
center_x, center_y = [geo.getFit2D()[k] for k in
['centerX', 'centerY']]
nx, ny = img.shape
mask_verts = [(center_x - bs_width, center_y),
(center_x, center_y - tri_offset),
(center_x + bs_width, center_y),
(center_x + bs_width + v_asym, ny),
(center_x - bs_width - v_asym, ny)]
x, y = np.meshgrid(np.arange(nx), np.arange(ny))
x, y = x.flatten(), y.flatten()
points = np.vstack((x, y)).T
path = Path(mask_verts)
grid = path.contains_points(points)
# Plug msk_grid into into next (edge-mask) step in automask
working_mask *= ~grid.reshape((ny, nx))
if alpha:
working_mask *= new_masking_method(img, geo, alpha=alpha,
tmsk=working_mask)
working_mask = working_mask.astype(np.bool)
return working_mask
def old_mask_img(img, geo,
edge=30,
lower_thresh=0.0,
upper_thresh=None,
bs_width=13, tri_offset=13, v_asym=0,
alpha=2.5,
tmsk=None):
"""
Mask an image based off of various methods
Parameters
----------
img: np.ndarray
The image to be masked
geo: pyFAI.geometry.Geometry
The pyFAI description of the detector orientation or any
subclass of pyFAI.geometry.Geometry class
edge: int, optional
The number of edge pixels to mask. Defaults to 30. If None, no edge
mask is applied
lower_thresh: float, optional
Pixels with values less than or equal to this threshold will be masked.
Defaults to 0.0. If None, no lower threshold mask is applied
upper_thresh: float, optional
Pixels with values greater than or equal to this threshold will be
masked.
Defaults to None. If None, no upper threshold mask is applied.
bs_width: int, optional
The width of the beamstop in pixels. Defaults to 13.
If None, no beamstop polygon mask is applied.
tri_offset: int, optional
The triangular pixel offset to create a pointed beamstop polygon mask.
Defaults to 13. If None, no beamstop polygon mask is applied.
v_asym: int, optional
The vertical asymmetry of the polygon beamstop mask. Defaults to 0.
If None, no beamstop polygon mask is applied.
alpha: float or tuple or, 1darray, optional
Then number of acceptable standard deviations, if tuple then we use
a linear distribution of alphas from alpha[0] to alpha[1], if array
then we just use that as the distribution of alphas. Defaults to 2.5.
If None, no outlier masking applied.
tmsk: np.ndarray, optional
The starting mask to be compounded on. Defaults to None. If None mask
generated from scratch.
Returns
-------
tmsk: np.ndarray
The mask as a boolean array. True pixels are good pixels, False pixels
are masked out.
"""
q = geo.qArray(img.shape)
qbinned = generate_binner(geo, img.shape)
if tmsk is None:
working_mask = np.ones(img.shape).astype(bool)
else:
working_mask = tmsk.copy()
if edge:
working_mask *= margin(img.shape, edge)
if lower_thresh:
working_mask *= (img >= lower_thresh).astype(bool)
if upper_thresh:
working_mask *= (img <= upper_thresh).astype(bool)
if all([a is not None for a in [bs_width, tri_offset, v_asym]]):
center_x, center_y = [geo.getFit2D()[k] for k in
['centerX', 'centerY']]
nx, ny = img.shape
mask_verts = [(center_x - bs_width, center_y),
(center_x, center_y - tri_offset),
(center_x + bs_width, center_y),
(center_x + bs_width + v_asym, ny),
(center_x - bs_width - v_asym, ny)]
x, y = np.meshgrid(np.arange(nx), np.arange(ny))
x, y = x.flatten(), y.flatten()
points = np.vstack((x, y)).T
path = Path(mask_verts)
grid = path.contains_points(points)
# Plug msk_grid into into next (edge-mask) step in automask
working_mask *= ~grid.reshape((ny, nx))
if alpha:
working_mask *= binned_outlier(img, q, alpha, qbinned.bin_edges,
mask=working_mask)
return working_mask
def mask_img(img, binner,
edge=30,
lower_thresh=0.0,
upper_thresh=None,
alpha=3,
auto_type='median',
tmsk=None):
"""
Mask an image based off of various methods
Parameters
----------
img: np.ndarray
The image to be masked
binner: pyFAI.geometry.Geometry
The pyFAI description of the detector orientation or any
subclass of pyFAI.geometry.Geometry class
edge: int, optional
The number of edge pixels to mask. Defaults to 30. If None, no edge
mask is applied
lower_thresh: float, optional
Pixels with values less than or equal to this threshold will be masked.
Defaults to 0.0. If None, no lower threshold mask is applied
upper_thresh: float, optional
Pixels with values greater than or equal to this threshold will be
masked.
Defaults to None. If None, no upper threshold mask is applied.
alpha: float, optional
Then number of acceptable standard deviations, if tuple then we use
a linear distribution of alphas from alpha[0] to alpha[1], if array
then we just use that as the distribution of alphas. Defaults to 3.
If None, no outlier masking applied.
auto_type: {'median', 'mean'}, optional
The type of binned outlier masking to be done, 'median' is faster,
where 'mean' is more accurate, defaults to 'median'.
tmsk: np.ndarray, optional
The starting mask to be compounded on. Defaults to None. If None mask
generated from scratch.
Returns
-------
tmsk: np.ndarray
The mask as a boolean array. True pixels are good pixels, False pixels
are masked out.
"""
if tmsk is None:
working_mask = np.ones(img.shape).astype(bool)
else:
working_mask = tmsk.copy()
if edge:
working_mask *= margin(img.shape, edge)
if lower_thresh:
working_mask *= (img >= lower_thresh).astype(bool)
if upper_thresh:
working_mask *= (img <= upper_thresh).astype(bool)
if alpha:
working_mask *= binned_outlier(img, binner, alpha=alpha,
tmsk=working_mask,
mask_method=auto_type)
working_mask = working_mask.astype(np.bool)
return working_mask
def mask_img(
img,
binner,
edge=30,
lower_thresh=0.0,
upper_thresh=None,
alpha=3,
auto_type="median",
tmsk=None,
pool=None,
):
"""
Mask an image based off of various methods
Parameters
----------
img: np.ndarray
The image to be masked
binner : BinnedStatistic1D instance
The binned statistics information
edge: int, optional
The number of edge pixels to mask. Defaults to 30. If None, no edge
mask is applied
lower_thresh: float, optional
Pixels with values less than or equal to this threshold will be masked.
Defaults to 0.0. If None, no lower threshold mask is applied
upper_thresh: float, optional
Pixels with values greater than or equal to this threshold will be
masked.
Defaults to None. If None, no upper threshold mask is applied.
alpha: float, optional
Then number of acceptable standard deviations, if tuple then we use
a linear distribution of alphas from alpha[0] to alpha[1], if array
then we just use that as the distribution of alphas. Defaults to 3.
If None, no outlier masking applied.
auto_type: {'median', 'mean'}, optional
The type of binned outlier masking to be done, 'median' is faster,
where 'mean' is more accurate, defaults to 'median'.
tmsk: np.ndarray, optional
The starting mask to be compounded on. Defaults to None. If None mask
generated from scratch.
pool : Executor instance
A pool against which jobs can be submitted for parallel processing
Returns
-------
tmsk: np.ndarray
The mask as a boolean array. True pixels are good pixels, False pixels
are masked out.
"""
if tmsk is None:
working_mask = np.ones(np.shape(img)).astype(bool)
else:
working_mask = tmsk.copy()
if edge:
working_mask *= margin(np.shape(img), edge)
if lower_thresh:
working_mask *= (img >= lower_thresh).astype(bool)
if upper_thresh:
working_mask *= (img <= upper_thresh).astype(bool)
if alpha:
working_mask *= binned_outlier(
img,
binner,
alpha=alpha,
tmsk=working_mask,
mask_method=auto_type,
pool=pool,
)
working_mask = working_mask.astype(np.bool)
return working_mask