forked from dstndstn/tractor
-
Notifications
You must be signed in to change notification settings - Fork 0
/
forced.py
152 lines (116 loc) · 4.2 KB
/
forced.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
import matplotlib
matplotlib.use('Agg')
import pylab as plt
import numpy as np
from tractor import *
from astrometry.util.plotutils import *
def main():
ps = PlotSequence('forced')
W,H = 50,50
sig1 = 1.
flux = 1000.
tim = Image(data=np.zeros((H,W)), inverr=np.ones((H,W)) / sig1,
psf=NCircularGaussianPSF([0.8], [1.]),
photocal=LinearPhotoCal(1.))
tim.sig1 = sig1
x1 = 20
src1 = PointSource(PixPos(x1, H/2), Flux(flux))
# What happens if we're missing one of the sources in the model?
for x2 in [30, 25, 24, 23, 22, 20]:
src2 = PointSource(PixPos(x2, H/2), Flux(flux))
fitsrc1 = PointSource(PixPos(x1, H/2), Flux(1.))
fitsrc1.freezeParam('pos')
(mod,modx,chix,flux0,fluxes) = runtest([tim], [src1, src2], [fitsrc1], 250)
plt.clf()
plt.imshow(mod, interpolation='nearest', origin='lower')
plt.title('Model: src2 distance = %i' % (x2 - x1))
ps.savefig()
plt.clf()
plt.imshow(modx, interpolation='nearest', origin='lower')
plt.title('Model: src2 distance = %i' % (x2 - x1))
ps.savefig()
plt.clf()
plt.imshow(chix, interpolation='nearest', origin='lower', vmin=-3, vmax=3)
plt.title('Chi: src2 distance = %i' % (x2 - x1))
ps.savefig()
xm = np.mean(fluxes)
st = np.std(fluxes)
xl = xm - 4.*st
xh = xm + 4.*st
plt.clf()
plt.hist(fluxes, bins=21, range=(xl,xh))
plt.axvline(flux0, color='r')
plt.axvline(flux, color='r', linestyle='--')
plt.xlim(xl,xh)
plt.xlabel('Fit Flux')
plt.title('Fits: src2 distance = %i' % (x2 - x1))
ps.savefig()
# What happens if we have the source position a little wrong?
chislices = []
for fx1 in [20, 21, 22, 23, 24]:
fitsrc1 = PointSource(PixPos(fx1, H/2), Flux(1.))
fitsrc1.freezeParam('pos')
(mod,modx,chix,flux0,fluxes) = runtest([tim], [src1], [fitsrc1], 250)
plt.clf()
plt.imshow(mod, interpolation='nearest', origin='lower')
plt.title('Model: pos error = %i' % (fx1 - x1))
ps.savefig()
plt.clf()
plt.imshow(modx, interpolation='nearest', origin='lower')
plt.title('Model: pos error = %i' % (fx1 - x1))
ps.savefig()
plt.clf()
plt.imshow(chix, interpolation='nearest', origin='lower', vmin=-3, vmax=3)
plt.title('Chi: pos error = %i' % (fx1 - x1))
ps.savefig()
chislices.append(chix[H/2, :])
xm = np.mean(fluxes)
st = np.std(fluxes)
xl = xm - 4.*st
xh = xm + 4.*st
plt.clf()
plt.hist(fluxes, bins=21, range=(xl,xh))
plt.axvline(flux0, color='r')
plt.axvline(flux, color='r', linestyle='--')
plt.xlim(xl,xh)
plt.xlabel('Fit Flux')
plt.title('Fits: pos error = %i' % (fx1 - x1))
ps.savefig()
plt.clf()
for chi in chislices:
plt.plot(chi)
plt.xlabel('Slice in x direction of image')
plt.ylabel('Chi of best fit')
ps.savefig()
def runtest(tims, realsrcs, fitsrcs, niters):
tr = Tractor(tims, realsrcs)
mod = tr.getModelImage(0)
tr = Tractor(tims, fitsrcs)
tr.freezeParam('images')
#tr.printThawedParams()
fluxes = []
for i in range(niters):
for tim in tims:
if i == 0:
noise = 0.
else:
noise = np.random.normal(scale=tim.sig1, size=tim.shape)
tim.data = mod + noise
if i == 1:
modx = tims[0].data
# while True:
# dlnp,X,alpha = tr.optimize(alphas=[0.1, 0.3, 1.])
# print 'dlnp', dlnp
# if dlnp < 0.1:
# break
tr.optimize_forced_photometry()
if i == 1:
chix = tr.getChiImage(0)
if i > 0:
fluxes.append(tr.getParams()[0])
else:
flux0 = tr.getParams()[0]
print 'Noise-free fit flux:', flux0
return mod, modx, chix, flux0, fluxes
if __name__ == '__main__':
main()