def setUp(self):
self.a = ImageArray(join(thisdir,
'coretestdata/im2_noannotations.png'))
self.a1 = ImageArray(join(thisdir, 'coretestdata/im1_annotated.png'))
self.a2 = STXMImage(
join(thisdir, "..", "..", "..", "sample-data",
"Sample_Image_2017-10-15_100.hdf5"))
self.a3 = STXMImage(
join(thisdir, "..", "..", "..", "sample-data",
"Sample_Image_2017-10-15_101.hdf5"))
class FuncsTest(unittest.TestCase):
def setUp(self):
self.a=ImageArray(join(thisdir,'coretestdata/im2_noannotations.png'))
self.a1=ImageArray(join(thisdir,'coretestdata/im1_annotated.png'))
self.a2=STXMImage(join(thisdir,"..","..","..","sample-data","Sample_Image_2017-10-15_100.hdf5"))
self.a3=STXMImage(join(thisdir,"..","..","..","sample-data","Sample_Image_2017-10-15_101.hdf5"))
def test_imagefile_ops(self):
self.a2.gridimage()
self.a3.gridimage()
self.a2.crop(5,-15,5,-5,_=True)
self.a3.crop(5,-15,5,-5,_=True)
self.b=self.a2//self.a3
self.assertEqual(self.b.shape,(90,80),"Failure to crop image correctly.")
self.assertGreater(self.b.max(),0.047,"XMCD Ratio calculation failed")
self.assertLess(self.b.min(),-0.05,"XMCD Ratio calculation failed")
self.b.normalise()
self.assertEqual(self.b.max(),1.0,"Normalise Image failed")
self.assertEqual(self.b.min(),-1.0,"Normalise Image failed")
self.profile=self.b.profile_line((0,0),(100,100))
self.profile.plot()
self.b.mask=self.a2.image>25E3
self.hist=self.b.hist(bins=200)
self.hist.column_headers=["XMCD Signal","Frequency"]
self.hist.labels=None
g1=LorentzianModel(prefix="g1_")
g2=LorentzianModel(prefix="g2_")
params=g1.make_params()
params.update(g2.make_params())
double_peak=g1+g2
g1=np.argmax(self.hist.y[:100]) # Location of first peak
g2=np.argmax(self.hist.y[100:])+100
for k, p in zip(params,[0.25,self.hist.x[g1],self.hist.y[g1]/np.sqrt(2),0.5,self.hist.y[g1],
0.25,self.hist.x[g2],self.hist.y[g2]/np.sqrt(2),0.5,self.hist.y[g2]]):
params[k].value=p
print(g1,g2,params)
self.res=self.hist.lmfit(double_peak,p0=params,output="report")
self.hist.add_column(self.res.init_fit,header="Initial Fit")
self.hist.add_column(self.res.best_fit,header="Best Fit")
self.hist.setas="xyyy"
self.hist.plot(fmt=["b+","b--","r-"])
plt.close("all")
#self.b.adnewjust_contrast((0.1,0.9),percent=True)z
def test_funcs(self):
b=self.a.translate((2.5,3))
self.c=b.correct_drift(ref=self.a)
self.d=b.align(self.a,method="scharr")
tv=np.array(self.d["tvec"])*10
cd=np.array(self.c["correct_drift"])*10
shift=np.array([25,30])
d1=mag(cd-shift)
d2=mag(tv-shift)
self.assertLess(d1,1.5,"Drift Correct off by more than 0.1 pxiels.")
self.assertLess(d2,1.5,"Scharr Alignment off by more than 0.1 pxiels.")
from os.path import join, dirname
import numpy as np
import matplotlib.pyplot as plt
from lmfit.models import LorentzianModel
from types import MethodType
# Load the images
thisdir = dirname(__file__)
imgs = ImageFolder(type=STXMImage)
for fname in [
"Sample_Image_2017-10-15_100.hdf5",
"Sample_Image_2017-10-15_101.hdf5",
]:
# Load the image
img = STXMImage(join(thisdir, "..", "..", "..", "sample-data", fname))
img.gridimage().crop(5, -15, 5, -5, _=True) # regularise grid and crop
imgs += img
# Align the two images
imgs.align(imgs[0], method="scharr", scale=4)
imgs.each.crop(2)
# Calculate the XMCD image and normalise it
xmcd = imgs[0] // imgs[1]
xmcd.normalise()
imgs += xmcd
strctural = (imgs[0] + imgs[1]) / 2
class FuncsTest(unittest.TestCase):
def setUp(self):
self.a=ImageArray(join(thisdir,'coretestdata/im2_noannotations.png'))
self.a1=ImageArray(join(thisdir,'coretestdata/im1_annotated.png'))
self.a2=STXMImage(join(thisdir,"..","..","..","sample-data","Sample_Image_2017-10-15_100.hdf5"))
self.a3=STXMImage(join(thisdir,"..","..","..","sample-data","Sample_Image_2017-10-15_101.hdf5"))
def test_imagefile_ops(self):
self.a2.gridimage()
self.a3.gridimage()
self.a2.crop(5,-15,5,-5,_=True)
self.a3.crop(5,-15,5,-5,_=True)
self.b=self.a2//self.a3
self.assertEqual(self.b.shape,(90,80),"Failure to crop image correctly.")
self.assertGreater(self.b.max(),0.047,"XMCD Ratio calculation failed")
self.assertLess(self.b.min(),-0.05,"XMCD Ratio calculation failed")
self.b.normalise()
self.assertEqual(self.b.max(),1.0,"Normalise Image failed")
self.assertEqual(self.b.min(),-1.0,"Normalise Image failed")
self.profile=self.b.profile_line((0,0),(100,100))
self.profile.plot()
self.b.mask=self.a2.image>25E3
self.hist=self.b.hist(bins=200)
self.hist.column_headers=["XMCD Signal","Frequency"]
self.hist.labels=None
g1=LorentzianModel(prefix="g1_")
g2=LorentzianModel(prefix="g2_")
params=g1.make_params()
params.update(g2.make_params())
double_peak=g1+g2
g1=np.argmax(self.hist.y[:100]) # Location of first peak
g2=np.argmax(self.hist.y[100:])+100
for k, p in zip(params,[0.25,self.hist.x[g1],self.hist.y[g1]/np.sqrt(2),0.5,self.hist.y[g1],
0.25,self.hist.x[g2],self.hist.y[g2]/np.sqrt(2),0.5,self.hist.y[g2]]):
params[k].value=p
print(g1,g2,params)
self.res=self.hist.lmfit(double_peak,p0=params,output="report")
self.hist.add_column(self.res.init_fit,header="Initial Fit")
self.hist.add_column(self.res.best_fit,header="Best Fit")
self.hist.setas="xyyy"
self.hist.plot(fmt=["b+","b--","r-"])
plt.close("all")
#self.b.adnewjust_contrast((0.1,0.9),percent=True)z
def test_funcs(self):
b=self.a.translate((2.5,3))
self.c=b.correct_drift(ref=self.a)
self.d=b.align(self.a,method="scharr")
tv=np.array(self.d["tvec"])*10
cd=np.array(self.c["correct_drift"])*10
shift=np.array([25,30])
d1=mag(cd-shift)
d2=mag(tv-(-shift[::-1]))
self.assertLess(d1,1.5,"Drift Correct off by more than 0.1 pxiels.")
self.assertLess(d2,1.5,"Scharr Alignment off by more than 0.1 pxiels.")
a1=ImageFile(self.a1.clone)
a1.as_float()
a1.image=np.sqrt(a1.image)/2+0.25
a1.adjust_contrast()
self.assertEqual(a1.span(),(0.0,1.0),"Either adjust_contrast or span failed with an ImageFile")
# print("#"*80)
# print(self.a.metadata)
# print(self.a1.metadata)
# print(all([k in self.a.metadata.keys() for k in self.a1.metadata.keys()]))
def test_imagefuncs(self):
self.a2.subtract_image(self.a2.image,offset=0)
self.assertTrue(np.all(self.a2.image<=0.0001),"Failed to subtract image from itself")
x=np.linspace(-3*np.pi,3*np.pi,101)
X,Y=np.meshgrid(x,x)
i=ImageFile(np.sin(X)*np.cos(Y))
i2=i.clone
j=i.fft()
self.assertTrue(np.all(np.unique(np.argmax(j,axis=1))==np.array([47,53])),"FFT of image test failed")
j.imshow()
self.assertTrue(len(plt.get_fignums())==1,"Imshow didn't open one window")
plt.close("all")
self.a2.imshow(title=None,figure=None)
self.a2.imshow(title="Hello",figure=1)
self.assertTrue(len(plt.get_fignums())==1,"Imshow with arguments didn't open one window")
plt.close("all")
i=i2
k=i+0.2*X-0.1*Y+0.2
k.level_image(mode="norm")
j=k-i
self.assertLess(np.max(j),0.01,"Level Image failed")
i2=i.clone
i2.quantize([-0.5,0,0.5])
self.assertTrue(np.all(np.unique(i2.data)==np.array([-0.5,0,0.5])),"Quantise levels failed")
i2=i.clone
i2.quantize([-0.5,0,0.5],levels=[-0.25,0.25])
self.assertTrue(np.all(np.unique(i2.data)==np.array([-0.5,0,0.5])),"Quantise levels failed")
i2=i.clone
i2.rotate(np.pi/4)
i2.fft()
self.assertTrue(np.all(np.unique(np.argmax(i2,axis=1))==np.array([46, 47, 49, 50, 51, 53])),"FFT of image test failed")
def setUp(self):
self.a=ImageArray(join(thisdir,'coretestdata/im2_noannotations.png'))
self.a1=ImageArray(join(thisdir,'coretestdata/im1_annotated.png'))
self.a2=STXMImage(join(thisdir,"..","..","..","sample-data","Sample_Image_2017-10-15_100.hdf5"))
self.a3=STXMImage(join(thisdir,"..","..","..","sample-data","Sample_Image_2017-10-15_101.hdf5"))
"""Demonstrate STXM Image Processing - G.Burnell Nov. 2017"""
from Stoner.Image import ImageFile, ImageFolder
from Stoner.HDF5 import STXMImage
from os.path import join, dirname
import numpy as np
import matplotlib.pyplot as plt
from lmfit.models import LorentzianModel
from types import MethodType
# Load the images
thisdir = dirname(__file__)
imgs = ImageFolder(type=STXMImage)
for fname in ["Sample_Image_2017-10-15_100.hdf5", "Sample_Image_2017-10-15_101.hdf5"]:
# Load the image
img = STXMImage(join(thisdir, "..", "..", "..", "sample-data", fname))
img.gridimage().crop(5, -15, 5, -5, _=True) # regularise grid and crop
imgs += img
# Align the two images
imgs[1].align(imgs[0], method="scharr")
# Calculate the XMCD image and normalise it
xmcd = imgs[0] // imgs[1]
xmcd.normalise()
imgs += xmcd
strctural = (imgs[0] + imgs[1]) / 2
imgs += strctural