def make_dark(fns, bad_pixels):
cfa0 = canont3i.read_CR2_as_CFA(fns[0])
height, width = cfa0.shape
stack = np.empty((height, width, len(fns)), dtype=cfa0.dtype)
for i, fn in enumerate(fns) :
print 'i: ', i
cfa = canont3i.read_CR2_as_CFA(fn)
stack[:,:,i] = cfa
print 'medianing'
med = np.median(stack)
print 'layer medianing'
dark = np.median(stack, axis=2)
for bp in bad_pixels:
dark[bp] = med
return dark
def find_badpixels(fns) :
bad_pixels = []
outliers = {}
for fn in fns :
print "Working on image: ", fn
cfa = canont3i.read_CR2_as_CFA(fn)
print "Image shape: ", cfa.shape
med = np.median(cfa)
print "Median: ", med
s = biweight.median_absolute_deviation(cfa, med)/0.6745
print "MAD: ", s*0.6745
print "StDev: ", s
image_outliers = np.argwhere(cfa > (med + 8*s))
image_outliers = np.concatenate([image_outliers, np.argwhere(cfa < (med - 8*s))])
print "Number of 8-sigma outliers:", image_outliers.shape[0]
print
print
for outlier in image_outliers:
key = tuple(outlier)
if key in outliers.keys():
outliers[key].append(fn)
else:
outliers[key] = [fn]
print "Bad pixels"
print "----------"
for pixel in outliers.keys():
if len(outliers[pixel]) > len(fns)**0.5:
print pixel, len(outliers[pixel])
bad_pixels.append(pixel)
return bad_pixels
