Remove bad images for one time correlation

Using same image (repeating the same image for 10000 times)



In [4]:

    
import numpy as np
from matplotlib.ticker import MaxNLocator
from matplotlib.colors import LogNorm
import matplotlib.pyplot as plt
import matplotlib.patches as mp
from matplotlib import cm
import matplotlib as mpl
%matplotlib notebook

import skbeam.core.correlation as corr
import skbeam.core.roi as roi
import skbeam.core.mask as mask



In [5]:

    
num_levels = 6
num_bufs = 8

# if we are 1.5 or later
if float('.'.join(mpl.__version__.split('.')[:2])) >= 1.5:
    cmap = 'viridis'
else:
    cmap = 'CMRmap'

# load the data
img_stack = np.load("100_500_NIPA_GEL.npy")
fig, ax = plt.subplots()
# plot the first image to make sure the data loaded correctly
ax.imshow(img_stack[0], cmap=cmap)
ax.set_title("NIPA_GEL_250K")









    














    











    Out[5]:





<matplotlib.text.Text at 0x103113128>



In [6]:

    
# define the ROIs
roi_start = 78 # in pixels
roi_width = 9 # in pixels
#roi_spacing = (5.0, 4.0)
x_center = 7. # in pixels
y_center = (129.) # in pixels
num_rings = 1

# get the edges of the rings
edges = roi.ring_edges(roi_start, width=roi_width, num_rings=num_rings)

# get the label array from the ring shaped 3 region of interests(ROI's)
labeled_roi_array = roi.rings(
    edges, (y_center, x_center), img_stack.shape[1:])



In [7]:

    
fig, ax = plt.subplots()
ax.imshow(labeled_roi_array)
plt.show()

Take the first image and repeat for 10000 times



In [9]:

    
y = []
for i in range(10000):
    y.append(img_stack[0])



In [10]:

    
g2y, lag_steps = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, np.asarray(y))

Results with no bad images



In [11]:

    
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "o", label="No bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.legend()
plt.show()

introduce 10 bad images



In [12]:

    
bad_list = np.random.choice(np.arange(10000), size=10)
imgs = mask.bad_to_nan_gen(y, bad_list)
bad_list









    Out[12]:





array([2024, 6672, 2297, 4944, 8489, 5627, 5102, 3799, 5945, 8224])



In [13]:

    
g2_10, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)



In [14]:

    
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps, g2_10[:, 0], "r*", label="with bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("10 bad images")
ax.legend()
plt.show()



In [ ]:

    
## 20 bad images



In [15]:

    
bad_list_2 = np.random.choice(np.arange(10000), size=20)

imgs = mask.bad_to_nan_gen(y, bad_list_2)
bad_list_2









    Out[15]:





array([6217, 3281, 2508, 6922, 4548, 8399, 3743, 1647, 8595, 1598, 3436,
       2369,  623, 4787, 7192, 6143, 5378, 4548, 3572, 5466])



In [16]:

    
g2_20, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)



In [17]:

    
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps, g2_20[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("20 bad images")
ax.legend()
plt.show()

50 bad images



In [18]:

    
bad_list_3 = np.random.choice(np.arange(10000), size=50)
bad_list_3









    Out[18]:





array([2440, 4951,  458, 3759, 1767, 1867, 1136, 7945, 4045,  184, 6226,
       7838, 6204, 2401, 3004, 9607, 8620, 6741, 5694, 9438, 5397, 3226,
       5653, 1235, 9841, 3192, 7777, 2084, 3632, 9950, 8007, 4776, 3935,
       8816, 5110, 4264, 8440, 2328, 3862, 9850, 5480, 7580, 9819, 8847,
       7202, 9054, 7851, 2630, 7061, 3972])



In [19]:

    
imgs = mask.bad_to_nan_gen(y, bad_list_3)
g2_50, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps, g2_50[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("50 bad images")
ax.legend()
plt.show()

100 bad images



In [20]:

    
bad_list_100 = np.random.choice(np.arange(10000), size=100)
bad_list_100









    Out[20]:





array([8057, 2089, 5277, 5526, 3021, 5268, 8970, 8380,  596,  344, 4046,
       9475, 2603, 1236, 3020, 1903, 4603, 9243, 1109,  743, 9848, 4311,
       5666, 5107, 4528, 3194, 2047, 2488, 5761, 6998, 7150, 9813, 2408,
       6594, 5395, 4040, 4611,  187, 5661, 2581, 7322, 8134, 2754, 9604,
       5936,  822, 9373, 6732,  608, 1697, 7269,  800, 7788, 9667, 3491,
       6900, 7955, 6057, 9149, 6623, 2472, 3237, 4399, 8354, 4330, 9197,
       4390, 8719, 4373, 9251, 8287, 9540,  548, 8662, 2403, 3812, 9242,
       6031, 9772, 6976, 5721, 2419, 9560, 4784, 5616, 3416,  349, 5058,
       9528, 3219, 9245, 4422, 4694, 6526, 6309, 6618, 1106, 7282, 3072,
       9454])



In [22]:

    
imgs = mask.bad_to_nan_gen(y, bad_list_100)
g2_100, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps, g2_100[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("100 bad images")
ax.legend()
plt.show()

200 bad images



In [23]:

    
bad_list_200 = np.random.choice(np.arange(10000), size=200)
bad_list_200









    Out[23]:





array([1758, 7053, 7748, 4720, 9488, 9246, 5457, 7837, 6840, 6527, 4553,
        300,  255, 5640, 1301, 2366, 9044, 7926, 5631,  135, 4672, 2094,
       4299, 6050, 2313, 2662, 3438, 8123, 2453, 9767, 1911, 3323, 4899,
       9736, 4204, 1907,  731, 1426, 2321, 4102, 9286, 5259, 1610, 5095,
       4416, 2057, 6798, 4832, 7640,  954, 8823, 1450, 2472, 5311, 6571,
       5025, 8899,  948, 2025, 8536, 1117, 4256, 6484, 9755, 7847, 8499,
       7211, 2870, 1750, 8128, 4135, 8545, 2682, 8729, 1460, 5809, 1079,
        940, 8751, 2415, 3446, 6324,  382,  307, 1522, 3171, 4002,  388,
        428, 1825, 4508, 3745, 4887, 9163, 6522, 4726, 5499, 6862, 6338,
        300, 6423,  415, 9123,    1,  729, 6496, 4678, 4270, 3009, 6186,
       7945, 4493, 2852, 3336, 5000, 8929, 6912, 5257, 6224, 2057, 2673,
       6341, 1866, 1725,  422,  629, 9893,  567, 7367, 2578, 8824, 5762,
       6868, 1183, 9085, 5629, 4854, 5528, 3177, 5232, 6730, 1279, 3995,
       5475, 1332, 6957, 8327, 5864, 9013, 3246, 9790, 1237, 9329, 2630,
       2996, 8053, 5655, 6656, 5714, 1031, 8807,  346, 1460, 7065, 7138,
       4782, 2422, 1540,  930, 1459, 3860, 1377, 1676, 5198, 6657, 9675,
        441, 2001, 6561, 3064, 3276, 5853, 9165, 7460, 1166, 5961,  264,
       8423, 9397, 2452, 4289, 5908, 2019, 5170, 4020, 3526,  711, 9360,
       6230, 5197])



In [24]:

    
imgs = mask.bad_to_nan_gen(y, bad_list_200)
g2_200, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps, g2_200[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("200 bad images")
ax.legend()
plt.show()

300 bad images



In [25]:

    
bad_list_300 = np.random.choice(np.arange(10000), size=300)
bad_list_300









    Out[25]:





array([4223, 5389, 1832, 2964, 9879, 6357, 9251, 1656, 8793, 7008,   92,
       1039, 3558, 5293, 6151, 6678, 4714, 1460, 3539, 5904, 2970, 7860,
       8292, 4329, 7511, 3706, 8104, 3904, 1881, 2355, 6105, 4925, 2297,
       3585, 5273, 9055, 7869, 5550, 5245,  572, 3580, 2070, 9942,  623,
       1924, 7270, 5843, 1356,  678, 9688, 7906, 1188, 5237, 5444, 2713,
       7375,  221, 4577, 5662, 6381, 4348, 2179, 9731, 7338, 1527, 3525,
       4324, 3992, 8985, 1704, 4849, 6841, 3129, 3224, 6866, 8742, 6291,
        591, 1068, 4894, 7784, 3702, 8062,  664, 8829, 1163, 4126, 3089,
       5047, 5421, 3788, 7449, 9650, 7379, 9430, 9408, 7446, 6928, 9763,
       5946, 3919, 5313, 1635, 2254, 1537, 2572, 8817, 3043, 6374, 7032,
       6563, 4776, 8206, 4904, 7347, 3831, 2184, 4876, 6379, 8518, 4058,
       6392, 3170, 9325, 9605, 9571, 5925, 6259, 7545, 9082, 1751, 2523,
       3503, 7187, 1883, 4526, 2818, 7173, 4753, 9589, 2433, 8944,  503,
       8963, 7252, 4334, 9837, 6035, 4554, 4472,  962, 4756, 9051,   53,
       5614, 4334, 8639, 8621, 1537, 7675, 6627, 5706, 4142, 2772, 4550,
       6888, 1030, 4681, 6772, 7819, 8617, 2697, 8946, 4065, 7332,  983,
       2079, 7991, 3699, 8586, 8218, 7348, 3473, 1235, 9949, 4674, 6832,
       4217, 6022, 3120, 9427, 5950, 4984, 3592,  266, 6065, 4610, 4321,
       8713, 4472,  384, 3074, 4894, 1069, 6541, 9252, 8325, 8913,  383,
       2458, 8613,  138, 3344, 3810, 3984, 1436, 7724, 6772, 3598, 1650,
       9567, 5420, 6074,  973, 9222, 2248, 8667, 5449, 7015, 4607, 2916,
       8795, 9869, 7844, 1525, 2369, 6803, 7339, 1044, 6605, 8002,  705,
       3591, 5712, 8923, 7876, 8142, 2606, 4112, 9433, 9483,  119, 2275,
       9131, 7827, 8670, 7592, 9336, 5298, 7005, 5827,  104, 1904, 1994,
       5965,  566,  770, 6970, 9652, 6299, 5670, 6337, 6932,  258, 6223,
       8732, 1611, 2550, 1600, 9407, 7477, 8582, 3111, 2302, 1762, 6078,
       7042, 7695, 8414,   53,  199, 7699, 2984, 5957, 6383, 9512, 4211,
       4293, 6955, 6401])



In [26]:

    
imgs = mask.bad_to_nan_gen(y, bad_list_300)
g2_300, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps, g2_300[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("300 bad images")
ax.legend()
plt.show()

500 bad images



In [28]:

    
bad_list_500 = np.random.choice(np.arange(10000), size=500)
imgs = mask.bad_to_nan_gen(y, bad_list_500)
g2_500, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps, g2_500[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("500 bad images")
ax.legend()
plt.show()

1000 bad images



In [31]:

    
bad_list_1000 = np.random.choice(np.arange(10000), size=1000)
imgs = mask.bad_to_nan_gen(y, bad_list_1000)
g2_1000, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps_n, g2_1000[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("1000 bad images")
ax.legend()
plt.show()

2000 bad images



In [32]:

    
bad_list_2000 = np.random.choice(np.arange(10000), size=2000)
imgs = mask.bad_to_nan_gen(y, bad_list_2000)
g2_2000, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps_n, g2_2000[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("5000 bad images")
ax.legend()
plt.show()

2500 bad images



In [33]:

    
bad_list_2500 = np.random.choice(np.arange(10000), size=2500)
imgs = mask.bad_to_nan_gen(y, bad_list_2500)
g2_2500, lag_steps_n = corr.multi_tau_auto_corr(num_levels, num_bufs,
                                        labeled_roi_array, imgs)
fig, ax = plt.subplots()
ax.semilogx(lag_steps, g2y[:, 0], "bo", label="No bad imgs")
ax.semilogx(lag_steps_n, g2_2500[:, 0], "r*", label="With bad imgs")
ax.set_ylim(bottom=1.014, top=1.018)
ax.set_title("2500 bad images")
ax.legend()
plt.show()



In [34]:

    
import skbeam



In [35]:

    
skbeam.__version__









    Out[35]:





'0.0.6+34.g73ae7f1'



In [ ]: