forked from revoltek/scripts
-
Notifications
You must be signed in to change notification settings - Fork 0
/
image_mosaic.py
executable file
·228 lines (200 loc) · 9.42 KB
/
image_mosaic.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
#!/usr/bin/env python
"""
A script to create a mosaic of LOFAR images
(The script was written to generate MSSS mosaics,
but it can be used in other cases as well)
HISTORY
=======
v0.1 G. Heald Created script
v0.2 R. Breton Fixed avgpb behavior and weighting implementation
v0.3 G. Heald Add weights, fix image naming, output sensitivity map
v0.4 S. van Velzen Add beam & frequency info to header
v0.5 S. van Velzen Use pyrap to save fits files
v0.6 J. Swinbank Select a single Stokes parameter from input maps
v0.7 G. Heald Fix RA behavior, add NCP option, pyfits tweak
v0.8 G. Heald Fix behavior near RA=0
v0.9 G. Heald Fix pyfits behavior (updating beam info)
TO DO
=====
Add NCP option (flag added in v0.7 but does not do anything)
"""
version = '0.9 2013-08-14'
import pyrap.images as pim
from pyrap import quanta
import numpy as np
import argparse
#import pylab as plt
import pyfits
import os
import time
def main(args):
# Generate lists of input images and check that they exist
images=[]
avgpbs=[]
psf_fwhm = [] # resolution
frequency = [] # frequency of images (should be equal?)
imagesbase=args.images.split(',')
for base in imagesbase:
images.append(base+'.'+args.extension)
if not os.path.exists(images[-1]):
print "Error: Image",images[-1],"does not exist"
return 1
avgpbs.append(base+'.'+args.avgpbext)
if not os.path.exists(avgpbs[-1]):
print "Error: PB image",avgpbs[-1],"does not exist"
return 1
# Collect weights and invert if requested
if args.weights == '':
weights = np.ones(len(images))
else:
weights = np.array(args.weights.split(',')).astype('float')
if args.invertwt:
weights = 1./weights
if len(weights) != len(images):
print "Error: List of weights is not the same length as list of images."
return 1
print "Combining images"
formstr = '{0:45s} {1:45s} {2:s} {3:s} {4:s} {5:s}'
print formstr.format("-----","--------","------------","-------","-------","------")
print formstr.format("Image", "PB image","Norm. weight", "Maj(ac)", "Min(ac)","PA(deg)")
print formstr.format("-----","--------","------------","-------","-------","------")
for i in range(len(images)):
this_pim = pim.image(images[i])
info_dict = this_pim.info()['imageinfo']['restoringbeam']
# get beam info
bpar_ma = quanta.quantity(info_dict['major']).get_value('deg')
bpar_mi = quanta.quantity(info_dict['minor']).get_value('deg')
bpar_pa = quanta.quantity(info_dict['positionangle']).get_value('deg')
psf_fwhm.append([bpar_ma, bpar_mi, bpar_pa])
frequency.append(this_pim.info()['coordinates']['spectral2']['restfreq'])
print '{0:45.45s} {1:45.45s} {2:0.2f} {3:0.2f} {4:0.2f} {5:0.2f}'.format(images[i], avgpbs[i], weights[i]/sum(weights), bpar_ma*60, bpar_mi*60,bpar_pa)
psf_fwhm = np.array(psf_fwhm)
frequency = np.array(frequency)
mean_psf_fwhm = np.mean(psf_fwhm, axis=0)
mean_frequency = np.mean(frequency)
print '\nmean Beam: {0:0.3f} maj (arcmin), {1:2.3f} min (arcmin), {2:0.2f} pa (deg)'.format(mean_psf_fwhm[0]*60, mean_psf_fwhm[1]*60, mean_psf_fwhm[2])
print '(Frequency (MHz):', mean_frequency*1e-6
if np.max(mean_frequency-frequency)/mean_frequency > 1e-6:
print '\n\nWARNING.\nAre you using images from different bands?'
print 'Frequencies (Hz):', frequency
# Initialize some vectors
declims = [] # store the limits of the declination axes
ralims = [] # store the limits of the r.a. axes
rainc = [] # store the r.a. increments in case they differ
decinc = [] # store the dec increments in case they differ
pims = [] # stores the pyrap images of the data
ppbs = [] # stores the pyrap images of the pb images
# Get image frames for input images
for im, pb in zip(images, avgpbs):
image = pim.image(im)
sptcoords = image.coordinates().get_coordinate('spectral')
nc = sptcoords.get_axis_size()
assert(sptcoords.get_image_axis() == 0)
# Get Stokes axis. Ensure we are working with the Stokes parameter requested.
stkcoords = image.coordinates().get_coordinate('stokes')
assert(stkcoords.get_image_axis() == 1)
if stkcoords.get_axis_size() == 1:
assert(stkcoords.get_stokes()[0] == args.stokes)
else:
stks = stkcoords.get_stokes().index(args.stokes)
image = image.subimage(blc=(0, stks), trc=(nc-1, stks), dropdegenerate=False)
ns = 1
dircoords = image.coordinates().get_coordinate('direction')
nx = dircoords.get_axis_size(axis=1)
ny = dircoords.get_axis_size(axis=0)
inc = dircoords.get_increment()
ref = dircoords.get_referencepixel()
val = dircoords.get_referencevalue()
ra_axis = (range(nx)-ref[1])*inc[1]+val[1]
dec_axis = (range(ny)-ref[0])*inc[0]+val[0]
rainc.append(inc[1])
decinc.append(inc[0])
declims.append(min(dec_axis))
declims.append(max(dec_axis))
mean_ra = np.mean(ra_axis)
ralims.append((min(ra_axis)-mean_ra)*np.cos(val[0])+mean_ra)
ralims.append((max(ra_axis)-mean_ra)*np.cos(val[0])+mean_ra)
pims.append(image)
ppbs.append(pim.image(pb))
# Generate the mosaic coordinate frame
master_dec = np.arange(min(declims),max(declims),min(decinc))
if max(ralims)-min(ralims) > 5.*np.pi/3.: # crossed RA=0
print "Warning: I think the mosaic crosses RA=0, treating the coordinates as such."
#ralims[ralims>np.pi] -= 2.*np.pi
for i in range(len(ralims)):
if ralims[i]>np.pi: ralims[i] = ralims[i]-2.*np.pi
master_ra = np.arange(max(ralims),min(ralims),max(rainc))
if args.verbose:
print "Found ra,dec pixel increments (arcsec):"
print np.array(rainc)*206265.
print np.array(decinc)*206265.
ma = pims[-1].coordinates()
ma['direction'].set_referencepixel([len(master_dec)/2,len(master_ra)/2])
ma['direction'].set_increment([min(decinc),max(rainc)])
ma['direction'].set_referencevalue([master_dec[len(master_dec)/2],master_ra[len(master_ra)/2]])
if args.NCP:
print 'Setting NCP projection is not yet working ....'
#ma['direction'].set_projection('ZEA')
# Initialize the arrays for the output image, sensitivity, and weights
master_im = np.zeros((len(master_dec),len(master_ra)))
master_weight = np.zeros((len(master_dec),len(master_ra)))
master_sens = np.zeros((len(master_dec),len(master_ra)))
# Reproject the images onto the master grid, weight and normalize
for i in range(len(pims)):
im = pims[i].regrid([2,3],ma,outshape=(nc,ns,len(master_dec),len(master_ra)))
pb = ppbs[i].regrid([2,3],ma,outshape=(nc,ns,len(master_dec),len(master_ra)))
imdata = np.squeeze(im.getdata())
pbdata = np.squeeze(pb.getdata())
newim = imdata
newpb = pbdata
newwt = (weights[i]*newpb)**2
master_im += newim*newwt
master_sens += newpb*newwt
master_weight += newwt
inds = master_weight != 0.
master_im[inds] /= master_weight[inds]
master_sens[inds] /= master_weight[inds]
# Show image if requested
if args.plotimg:
plt.imshow(master_im,vmin=0.,vmax=0.5)
plt.show()
# Write fits files
arrax = np.zeros( (1,1, len(master_im[:,0]), len(master_im[0,:])) )
arrax[0,0,:,:] = master_im
# Open new casa image for mosaic
new_pim = pim.image('',shape=(1,1, len(master_dec),len(master_ra)), coordsys=ma)
new_pim.putdata(arrax)
# Write fits
new_pim.tofits(args.outfits, overwrite=True)
# Same for sensitivity
new_pim_sens = pim.image('',shape=(1,1,len(master_dec),len(master_ra)),coordsys=ma)
arrax[0,0,:,:] = master_sens
new_pim_sens.putdata(arrax) #
new_pim_sens.tofits(args.sensfits, overwrite=True)
# need to add new beam info (not sure if this is possible with pyrap)
hdu = pyfits.open(args.outfits,mode='update')
header = hdu[0].header
header.update('BMAJ',mean_psf_fwhm[0])
header.update('BMIN',mean_psf_fwhm[1])
header.update('BPA',mean_psf_fwhm[2])
header.update('BUNIT',pims[-1].info()['unit'])
header.update('RESTFRQ',mean_frequency)
header.update('RESTFREQ',mean_frequency)
newhdu = pyfits.PrimaryHDU(data=hdu[0].data, header=header)
newhdu.writeto(args.outfits,clobber=True)
return
print "LOFAR mosaic generator, v"+version+'\n'
parser = argparse.ArgumentParser(description="Mosaic MSSS images.")
parser.add_argument('images',metavar='list_of_images',help='Input image names without extension, as a comma separated list (no spaces). Note that the "0" in the avgpb filenames will be accounted for in the script, do not include it here.')
parser.add_argument('-x','--extension',help='Image extension to combine [default restored.corr]',default='restored.corr')
parser.add_argument('-a','--avgpbext',help='Extension for primary beam images [default avgpb]',default='avgpb')
parser.add_argument('-w','--weights',help='Image weights, as a comma separated list (this must be of the same length and in same order as the IMAGES list) [default equal weights]',default='')
parser.add_argument('-n','--invertwt',help='Invert weights before applying? (Can be useful if image noise values are given as input to WEIGHTS) [default False]',action='store_true',default=False)
parser.add_argument('-v','--verbose',help='Give some verbose output [default False]',action='store_true',default=False)
parser.add_argument('-o','--outfits',help='Output name of mosaic fits file [default mosaic.fits]',default='mosaic.fits')
parser.add_argument('-s','--sensfits',help='Output name of sensitivity fits file [default sensitivity.fits]',default='sensitivity.fits')
parser.add_argument('-p','--plotimg',help='Display image on screen? [default False]',action='store_true',default=False)
parser.add_argument('-S','--stokes',help='Stokes parameter to use? [default I]',default='I')
parser.add_argument('-N','--NCP',help='Use NCP instead of SIN? This option does not work yet. [default False]',default=False,action='store_true')
args = parser.parse_args()
main(args)