Différences

Ci-dessous, les différences entre deux révisions de la page.

--- recherche:methodes:ebsd:edax_conv [27/10/2020 09:47] – créée simon.breumier
+++ recherche:methodes:ebsd:edax_conv [27/10/2020 10:16] (Version actuelle) – simon.breumier
@@ Ligne 1: / Ligne 1: @@
 ==== Convert Oxford EBSD patterns in HDF5 for EMSOFT ====
+The following python script converts a list of pattern in the //picRep// folder with naming conventions "NAME_ID" where ID = Nx*y+x, into an HDF5 format compatible with EMSOFT.
+To use it, change the values of the variables in the PARAMETERS section to the one that fits to you.
+Note that you will need an HDF5 template file, I usually use one of the HDF5 provided for the [[https://kilthub.cmu.edu/articles/Data_Sets_and_Analysis_Results_Accompanying_a_Dictionary_Indexing_Tutorial_Paper/7792505|EDAX Nickel example]] on the EMSOFT website:
+<Code:python>
+##########################################################################
+# Converts a set of Kikuchi pattern to a readable EDAX HDF5 format
+# Kikuchi pattern must all be in the same folder (defined by 'picRep' variable)
+# Pattern names must be of the form 'Something_ID', where ID = Nx*y+x
+# As all the patterns are stored in RAM for writing, it is possible to make an HDF5
+# of a subdomain of the map
+#
+# Note: for now, the scrip just fills the Patterns and their coordinates on the map
+# All other fields contain dummy values.
+#
+# 31/10/2019			Simon Breumier			simon.breumier@mines-saint-etienne.org
+##########################################################################
+import h5py
+import matplotlib.pyplot as plt
+import matplotlib.image as mpimg
+from PIL import Image
+import numpy as np
+import os
+from scipy import ndimage
+import math
+from get_tiff_param import get_pat_pc
+from progress.bar import Bar
+print('******************** Pattern picture ==> EDAX HDF5 conversion *********')
+####################	P	A	R	A	M	E	T	E	R	S	###########
+out = 'MTL_100_bin2' # output file name
+picRep = 'GND_100_CP/Pattern' #pattern pic repository path
+template_file='EDAX-Ni.h5' #HDF5 template file name
+# Carto parameters
+Xsize=201 #Map width in px
+Ysize=164 #Map height in px
+X0 = 1 #X position to start the writing (usefull to make an HDF5 file with a partition of your map)
+Y0 = 1 #Y position to start the writing
+Xmax = 201 #X position to stop the writting
+Ymax = 164 #Y position to stop the writting
+step=0.003 #Step size in mm
+#Pattern dimensions (after binning)
+pattHeight = 512
+pattWidth = 672
+#####################################
+tmp = h5py.File(template_file,'r')
+f = h5py.File(out+'.h5','w')
+for elem in tmp.keys():
+	tmp.copy(elem,f)
+tmp.close()
+#############	H	E	A	D	E	R
+del(f[' Manufacturer'])
+f.create_dataset(' Manufacturer', data=np.array([b'Oxford']))
+head = f['Scan 1']['EBSD']['Header']
+print('** Fill Header data')
+tmp = f['Scan 1']['EBSD']['Header']['Pattern Height']
+tmp[...] = pattHeight
+tmp = f['Scan 1']['EBSD']['Header']['Pattern Width']
+tmp[...] = pattWidth
+tmp = f['Scan 1']['EBSD']['Header']['Sample Tilt']
+tmp[...] = 70.
+tmp = f['Scan 1']['EBSD']['Header']['Step X']
+tmp[...] = step*1000
+tmp = f['Scan 1']['EBSD']['Header']['Step Y']
+tmp[...] = step*1000
+tmp = f['Scan 1']['EBSD']['Header']['Working Distance']
+tmp[...] = 16022.
+tmp = f['Scan 1']['EBSD']['Header']['nColumns']
+tmp[...] = Xmax - X0+1
+tmp = f['Scan 1']['EBSD']['Header']['nRows']
+tmp[...] = Ymax - Y0+1
+#############	E	B	S	D	 	D	A	T	A
+if 'Scan 4' in f.keys():
+	del(f['Scan 4'])
+if 'Scan 6' in f.keys():
+	del(f['Scan 6'])
+ebsdDat =f['Scan 1']['EBSD']
+print('** Fill EBSD data')
+print('listing EBSD patterns...')
+listPat = os.listdir(picRep)
+if 'Thumbs.db' in listPat: listPat.remove('Thumbs.db')
+lenPat = len(listPat)
+print('Filling dummy values...')
+dum = np.ones((Xsize-X0)*(Ysize-Y0)+1)
+i =0
+if 'X Position' in f['Scan 1']['EBSD']['Data'].keys():
+	del(f['Scan 1']['EBSD']['Data']['X Position'])
+if 'Y Position' in f['Scan 1']['EBSD']['Data'].keys():
+	del(f['Scan 1']['EBSD']['Data']['Y Position'])
+if 'Pattern' in f['Scan 1']['EBSD']['Data'].keys():
+	del(f['Scan 1']['EBSD']['Data']['Pattern'])
+dumList = ['CI','Fit','IQ','PRIAS Bottom Strip','PRIAS Center Square','PRIAS Top Strip','Phase','Phi','Phi1','Phi2','SEM Signal','Valid']
+for elem in dumList:
+	if elem in f['Scan 1']['EBSD']['Data'].keys():
+		del(f['Scan 1']['EBSD']['Data'][elem])
+	#f['Scan 1']['EBSD']['Data'].create_dataset(elem,data=dum)
+#############
+lenTot = (Xmax-X0+1)*(Ymax-Y0+1)
+f['Scan 1']['EBSD']['Data'].create_dataset("X Position",data=np.zeros(lenTot))
+f['Scan 1']['EBSD']['Data'].create_dataset("Y Position",data=np.zeros(lenTot))
+f['Scan 1']['EBSD']['Data'].create_dataset("X BEAM",data=np.zeros(lenTot))
+f['Scan 1']['EBSD']['Data'].create_dataset("Y BEAM",data=np.zeros(lenTot))
+f['Scan 1']['EBSD']['Data'].create_dataset("Pattern",(lenTot,pattHeight,pattWidth), dtype="u1")
+Patval = []
+DDs = []
+PCXs = []
+PCYs = []
+i = 0
+last = 0
+j = 0
+limAct = [Ysize/4,Ysize/2,3*Ysize/4,Ysize]
+tmp,tmp2,picId = listPat[2].split('_')
+tmp = tmp+'_'+tmp2
+numZeros = len(picId.split('.')[0])-1
+print('Find initial point')
+bar = Bar('Processing', max=(Xmax-X0+1)*(Ymax-Y0+1))
+for Yact in range(Y0,Ymax+1):
+	for Xact in range(X0,Xmax+1):
+		picId = (Yact-1)*Xsize+Xact
+		picId_str = (numZeros-math.floor(math.log10(picId)))*'0'+str(picId)
+		PCX, PCY, lfov, Xbeam, Ybeam, DD = get_pat_pc(picRep+'/'+tmp+'_'+picId_str+'.tiff')
+		tabID = (Yact-Y0)*(Xmax-X0+1)+Xact-X0
+		f['Scan 1']['EBSD']['Data']['X BEAM'][tabID] = Xact-X0
+		f['Scan 1']['EBSD']['Data']['Y BEAM'][tabID] = Yact-Y0
+		DDs.append(DD*pattWidth/pattHeight)
+		PCXs.append(PCX)
+		PCYs.append(PCY)
+		imActu = mpimg.imread(picRep+'/'+tmp+'_'+picId_str+'.tiff')
+		f['Scan 1']['EBSD']['Data']['X Position'][tabID] = (Xact-X0)*step
+		f['Scan 1']['EBSD']['Data']['Y Position'][tabID] = (Yact-Y0)*step
+		f['Scan 1']['EBSD']['Data']['Pattern'][tabID,:,:] = imActu
+		i+=1
+		bar.next()
+bar.finish()
+print('\n saving ...')
+f['Scan 1']['EBSD']['Data'].create_dataset("Phase",data=np.ones(((Xmax-X0+1)*(Ymax-Y0+1))))
+f['Scan 1']['EBSD']['Data'].create_dataset("DD",data=np.asarray(DDs))
+f['Scan 1']['EBSD']['Data'].create_dataset("PCX",data=np.asarray(PCXs))
+f['Scan 1']['EBSD']['Data'].create_dataset("PCY",data=np.asarray(PCYs))
+f.close()
+print('done!')
+print('*************************************************')
+</Code>