__author__ = "sibirrer"
import numpy as np
from scipy import ndimage
from scipy import interpolate
import copy
import lenstronomy.Util.util as util
from lenstronomy.Util.package_util import exporter
export, __all__ = exporter()
[docs]
@export
def add_layer2image(grid2d, x_pos, y_pos, kernel, order=1):
"""Adds a kernel on the grid2d image at position x_pos, y_pos with an interpolated
subgrid pixel shift of order=order.
:param grid2d: 2d pixel grid (i.e. image)
:param x_pos: x-position center (pixel coordinate) of the layer to be added
:param y_pos: y-position center (pixel coordinate) of the layer to be added
:param kernel: the layer to be added to the image
:param order: interpolation order for sub-pixel shift of the kernel to be added
:return: image with added layer, cut to original size
"""
x_int = int(round(x_pos))
y_int = int(round(y_pos))
shift_x = x_int - x_pos
shift_y = y_int - y_pos
kernel_shifted = ndimage.shift(kernel, shift=[-shift_y, -shift_x], order=order)
return add_layer2image_int(grid2d, x_int, y_int, kernel_shifted)
[docs]
@export
def add_layer2image_int(grid2d, x_pos, y_pos, kernel):
"""Adds a kernel on the grid2d image at position x_pos, y_pos at integer positions
of pixel.
:param grid2d: 2d pixel grid (i.e. image)
:param x_pos: x-position center (pixel coordinate) of the layer to be added
:param y_pos: y-position center (pixel coordinate) of the layer to be added
:param kernel: the layer to be added to the image
:return: image with added layer
"""
nx, ny = np.shape(kernel)
if nx % 2 == 0:
raise ValueError("kernel needs odd numbers of pixels")
num_x, num_y = np.shape(grid2d)
x_int = int(round(x_pos))
y_int = int(round(y_pos))
k_x, k_y = np.shape(kernel)
k_l2_x = int((k_x - 1) / 2)
k_l2_y = int((k_y - 1) / 2)
min_x = np.maximum(0, x_int - k_l2_x)
min_y = np.maximum(0, y_int - k_l2_y)
max_x = np.minimum(num_x, x_int + k_l2_x + 1)
max_y = np.minimum(num_y, y_int + k_l2_y + 1)
min_xk = np.maximum(0, -x_int + k_l2_x)
min_yk = np.maximum(0, -y_int + k_l2_y)
max_xk = np.minimum(k_x, -x_int + k_l2_x + num_x)
max_yk = np.minimum(k_y, -y_int + k_l2_y + num_y)
if (
min_x >= max_x
or min_y >= max_y
or min_xk >= max_xk
or min_yk >= max_yk
or (max_x - min_x != max_xk - min_xk)
or (max_y - min_y != max_yk - min_yk)
):
return grid2d
kernel_re_sized = kernel[min_yk:max_yk, min_xk:max_xk]
new = grid2d.copy()
new[min_y:max_y, min_x:max_x] += kernel_re_sized
return new
[docs]
@export
def add_background(image, sigma_bkd):
"""Generates background noise to image. To generate a noisy image with background
noise, generate image_noisy = image + add_background(image, sigma_bkd)
:param image: pixel values of image
:param sigma_bkd: background noise (sigma)
:return: a realisation of Gaussian noise of the same size as image
"""
nx, ny = np.shape(image)
background = np.random.randn(nx, ny) * sigma_bkd
return background
[docs]
@export
def add_poisson(image, exp_time):
"""Generates a poison (or Gaussian) distributed noise with mean given by surface
brightness. To generate a noisy image with Poisson noise, perform image_noisy =
image + add_poisson(image, exp_time)
:param image: pixel values (photon counts per unit exposure time)
:param exp_time: exposure time
:return: Poisson noise realization of input image
"""
# Gaussian approximation for Poisson distribution, normalized to exposure time
sigma = np.sqrt(np.abs(image) / exp_time)
nx, ny = np.shape(image)
poisson = np.random.randn(nx, ny) * sigma
return poisson
[docs]
@export
def rotateImage(img, angle):
"""Querries scipy.ndimage.rotate routine :param img: image to be rotated :param
angle: angle to be rotated (radian) :return: rotated image."""
imgR = ndimage.rotate(img, angle, reshape=False)
return imgR
[docs]
@export
def shift_image(img, shift):
"""Queries scipy.ndimage.shift routine.
:param img: image to be shifted
:param shift: sequence containing x and y shift in pixels
:return: shifted image
"""
img_s = ndimage.shift(img, shift)
return img_s
[docs]
@export
def re_size_array(x_in, y_in, input_values, x_out, y_out):
"""Resizes 2d array (i.e. image) to new coordinates. So far only works with square
output aligned with coordinate axis.
:param x_in:
:param y_in:
:param input_values:
:param x_out:
:param y_out:
:return:
"""
# from skimage.transform import resize
# resize(input_values)
# interp_2d = interpolate.interp2d(x_in, y_in, input_values, kind="linear")
# out_values = interp_2d.__call__(x_out, y_out)
from scipy.interpolate import RectBivariateSpline
func = RectBivariateSpline(x_in, y_in, z=input_values, kx=1, ky=1, s=0)
return func(x_out, y_out)
[docs]
@export
def symmetry_average(image, symmetry):
"""Symmetry averaged image.
:param image:
:param symmetry:
:return:
"""
img_sym = np.zeros_like(image)
angle = 360.0 / symmetry
for i in range(symmetry):
img_sym += rotateImage(image, angle * i)
img_sym /= symmetry
return img_sym
[docs]
@export
def findOverlap(x_mins, y_mins, min_distance):
"""Finds overlapping solutions, deletes multiples and deletes non-solutions and if
it is not a solution, deleted as well."""
n = len(x_mins)
idex = []
for i in range(n):
if i == 0:
pass
else:
for j in range(0, i):
if (
abs(x_mins[i] - x_mins[j]) < min_distance
and abs(y_mins[i] - y_mins[j]) < min_distance
):
idex.append(i)
break
x_mins = np.delete(x_mins, idex, axis=0)
y_mins = np.delete(y_mins, idex, axis=0)
return x_mins, y_mins
[docs]
@export
def coordInImage(x_coord, y_coord, num_pix, deltapix):
"""
checks whether image positions are within the pixel image in units of arcsec
if not: remove it
:returns: image positions within the pixel image
"""
idex = []
min_ = -deltapix * num_pix / 2
max_ = deltapix * num_pix / 2
for i in range(len(x_coord)): # sum over image positions
if (
x_coord[i] < min_
or x_coord[i] > max_
or y_coord[i] < min_
or y_coord[i] > max_
):
idex.append(i)
x_coord = np.delete(x_coord, idex, axis=0)
y_coord = np.delete(y_coord, idex, axis=0)
return x_coord, y_coord
[docs]
@export
def re_size(image, factor=1):
"""Re-sizes image with nx x ny to nx/factor x ny/factor.
:param image: 2d image with shape (nx,ny)
:param factor: integer >=1
:return:
"""
if factor < 1:
raise ValueError("scaling factor in re-sizing %s < 1" % factor)
elif factor == 1:
return image
f = int(factor)
nx, ny = np.shape(image)
if int(nx / f) == nx / f and int(ny / f) == ny / f:
small = image.reshape([int(nx / f), f, int(ny / f), f]).mean(3).mean(1)
return small
else:
raise ValueError(
"scaling with factor %s is not possible with grid size %s, %s" % (f, nx, ny)
)
[docs]
@export
def rebin_image(bin_size, image, wht_map, sigma_bkg, ra_coords, dec_coords, idex_mask):
"""Re-bins pixels, updates cutout image, wht_map, sigma_bkg, coordinates, PSF.
:param bin_size: number of pixels (per axis) to merge
:return:
"""
numPix = int(len(image) / bin_size)
numPix_precut = numPix * bin_size
factor = int(len(image) / numPix)
if not numPix * bin_size == len(image):
image_precut = image[0:numPix_precut, 0:numPix_precut]
else:
image_precut = image
image_resized = re_size(image_precut, factor)
image_resized *= bin_size**2
wht_map_resized = re_size(wht_map[0:numPix_precut, 0:numPix_precut], factor)
sigma_bkg_resized = bin_size * sigma_bkg
ra_coords_resized = re_size(ra_coords[0:numPix_precut, 0:numPix_precut], factor)
dec_coords_resized = re_size(dec_coords[0:numPix_precut, 0:numPix_precut], factor)
idex_mask_resized = re_size(idex_mask[0:numPix_precut, 0:numPix_precut], factor)
idex_mask_resized[idex_mask_resized > 0] = 1
return (
image_resized,
wht_map_resized,
sigma_bkg_resized,
ra_coords_resized,
dec_coords_resized,
idex_mask_resized,
)
[docs]
@export
def stack_images(image_list, wht_list, sigma_list):
"""Stacks images and saves new image as a fits file.
:return:
"""
image_stacked = np.zeros_like(image_list[0])
wht_stacked = np.zeros_like(image_stacked)
sigma_stacked = 0.0
for i in range(len(image_list)):
image_stacked += image_list[i] * wht_list[i]
sigma_stacked += sigma_list[i] ** 2 * np.median(wht_list[i])
wht_stacked += wht_list[i]
image_stacked /= wht_stacked
sigma_stacked /= np.median(wht_stacked)
wht_stacked /= len(wht_list)
return image_stacked, wht_stacked, np.sqrt(sigma_stacked)
[docs]
@export
def cut_edges(image, num_pix):
"""Cuts out the edges of a 2d image and returns re-sized image to numPix center is
well defined for odd pixel sizes.
:param image: 2d numpy array
:param num_pix: square size of cut out image
:return: cutout image with size numPix
"""
nx, ny = image.shape
if nx < num_pix or ny < num_pix:
raise ValueError(
"image can not be resized, in routine cut_edges with image shape (%s %s) "
"and desired new shape (%s %s)" % (nx, ny, num_pix, num_pix)
)
if (nx % 2 == 0 and ny % 2 == 1) or (nx % 2 == 1 and ny % 2 == 0):
raise ValueError(
"image with odd and even axis (%s %s) not supported for re-sizing"
% (nx, ny)
)
if (nx % 2 == 0 and num_pix % 2 == 1) or (nx % 2 == 1 and num_pix % 2 == 0):
raise ValueError(
"image can only be re-sized from even to even or odd to odd number."
)
x_min = int((nx - num_pix) / 2)
y_min = int((ny - num_pix) / 2)
x_max = nx - x_min
y_max = ny - y_min
resized = image[x_min:x_max, y_min:y_max]
return copy.deepcopy(resized)
[docs]
@export
def radial_profile(data, center):
"""Computes radial profile.
:param data: 2d numpy array
:param center: center [x, y] from which pixel to compute the radial profile
:return: radial profile (in units pixel)
"""
y, x = np.indices(data.shape)
r = np.sqrt((x - center[0]) ** 2 + (y - center[1]) ** 2)
r = r.astype(int)
tbin = np.bincount(r.ravel(), data.ravel())
nr = np.bincount(r.ravel())
radialprofile = tbin / nr
return radialprofile
[docs]
@export
def gradient_map(image):
"""Computes gradients of images with the sobel transform.
:param image: 2d numpy array
:return: array of same size as input, with gradients between neighboring pixels
"""
from skimage import filters
return filters.sobel(image)