def getOffset(self, position):
ret = {"X": 0.0, "Y": 0.0, "N": 0.0, "E": 0.0, "Status": self.state()}
try:
frame = self._takeImage()
fname = "/tmp/autoguider.fits"
if os.path.exists(fname):
os.remove(fname)
frame.save(fname)
img = fits.getdata(fname)
# Extract some backgroud
img -= np.mean(img) * 0.9
img[img < 0] = 0.0
pY, pX = centroid(img)
ret["X"] = pX - position["XWIN_IMAGE"]
ret["Y"] = pY - position["YWIN_IMAGE"]
centerPos = frame.worldAt([position["XWIN_IMAGE"], position["YWIN_IMAGE"]])
currPos = frame.worldAt([pX, pY])
# offset = centerPos.angsep(currPos)
ret["E"] = Coord.fromAS((centerPos.ra.AS - currPos.ra.AS) * np.cos(currPos.dec.R))
ret["N"] = Coord.fromAS(centerPos.dec.AS - currPos.dec.AS)
except:
if self.abort.isSet():
ret["Status"] = GuiderStatus.ABORTED
return ret
else:
raise
self.plot(frame, position, ret)
return ret
def getCOM(self, gf):
"""
Calculates and stores the centroid of each star passed
in as a numpy array.
:param gf: fits file of single star, sized as guider box.
"""
# New version, uses astropysics.util.alg func.
# Ego safety note: the results are within 0.01% var.
# from the old method...
self.centroids.append(centroid(gf))
def main(argv):
from optparse import OptionParser
parser = OptionParser()
parser.add_option('-f', '--filename',
help='Input list of fits files.'
, type='string')
parser.add_option('-v', '--verbose',
help='Run in verbose mode.', action='store_true',
default=False)
opt, args = parser.parse_args(argv)
if opt.verbose:
log.setLevel(logging.DEBUG)
# 1 - Read list of images
imglist = np.loadtxt(opt.filename,dtype='S')
# 2 - First image is reference
refimg = fits.open(imglist[0])
# 2 - Telescope characteristics
telescope_platescale = 133 # arcsec/mm
# 2 - Autoguider center
# For an SBIG ST7XME the parameters of the offset between the CCD and the guider are:
guider_offset = {'x': 0, 'y': 0} # in mm
guider_size = {'x': refimg[0].header['CCDPXSZX']*1e-3*refimg[0].header['CCD_DIMX'],
'y': refimg[0].header['CCDPXSZY']*1e-3*refimg[0].header['CCD_DIMY'],
'x_px': refimg[0].header['CCD_DIMX'],
'y_px': refimg[0].header['CCD_DIMY']} # in mm / px and pixels
guider_pixelsize = {'x': refimg[0].header['CCDPXSZX']*1e-3,
'y': refimg[0].header['CCDPXSZY']*1e-3} # in mm
guider_pixelsize.update({'x_ang': telescope_platescale * guider_pixelsize['x'],
'y_ang': telescope_platescale * guider_pixelsize['y']}) # in arcsec / px
# For plotting purposes
ccd_size = {'x': guider_size['x'], 'y': guider_size['y']} # in mm
ccd_pixelsize = {'x': guider_pixelsize['x'], 'y': guider_pixelsize['y']} # in mm
# Get a DSS Image...
img2 = np.array(refimg[0].data,copy=True)
img = np.array(refimg[0].data,copy=True)
# Image sizes in pixels
imsize = int(round(guider_size['x'] / guider_pixelsize['x'])), int(round(guider_size['y'] / guider_pixelsize['y']))
# img = imresize(img, imsize)
# Image sizes in arcmins
aux_width = guider_size['x'] * telescope_platescale / 60
aux_height = guider_size['y'] * telescope_platescale / 60
# TODO: DELME
img_scale = 1.2 # in arcsec / pix
imsize = img.shape[0], img.shape[1]
imsize_resampled = int(round(imsize[0] * img_scale / guider_pixelsize['x_ang'])), \
int(round(imsize[1] * img_scale / guider_pixelsize['y_ang']))
pix = imresize(img, imsize_resampled)
# TODO: DELME END
px_border = 20 # Number of pixels to throw away due to border-effect
max_offset = 10 # Maximum offset between images, in pixels
img_corner = [int(round(pix.shape[0] / 2 - guider_size['y_px'] / 2)),
int(round(pix.shape[1] / 2 - guider_size['x_px'] / 2))]
mask = img < 1.1 * np.median(img) # Background / weak stars subtraction
#print img_new
img[mask] = 0.
segmap = label(img > 0)
print segmap[segmap>0]
# fig2 = plt.figure(2)
# plt.clf()
# plt.imshow(segmap)
plt.figure(1)
cent = []
for i_object in range(1, np.max(segmap) + 1):
mm = segmap == i_object
if len(mm[mm]) > 1:
cent.append(centroid(img * mm))
# plt.plot(cent_new[-1][1], cent_new[-1][0], '.', lw=2, color='black')
# if i > 0:
# plt.plot(cent[-1][1], cent[-1][0], '.', lw=2, color='black')
cent = np.array(cent)
print img2.min(),img2.max(),img2.mean()
plt.imshow(img2,cmap=plt.cm.gray,interpolation='nearest',origin='lower',
vmin=img2.min(),
vmax=img2.mean()*1.1)
for i in range(len(cent)):
plt.plot(cent[i][1],cent[i][0],marker='o',markerfacecolor='none',mec='r')
plt.xlim(0, img.shape[1])
plt.ylim(0, img.shape[0])
# plt.show()
# return 0
for i in range(len(imglist)):
#offset = np.random.randint(-10, 10, size=2)
#img_corner = np.sum([img_corner, offset], axis=0) # Add an offset to the corner pixel of the image
#img_cut_new = np.copy(
# pix[img_corner[0]:img_corner[0] + guider_size['y_px'], img_corner[1]:img_corner[1] + guider_size['x_px']])
img_cut_new = fits.getdata(imglist[i])
img_cut_new = img_cut_new[px_border:-px_border, px_border:-px_border]
fig1 = plt.figure(1)
plt.clf()
plt.imshow(img_cut_new, ,interpolation='nearest',origin='lower',
vmin=img2.min(),
vmax=img2.mean()*1.1)
plt.xlim(0, img_cut_new.shape[1])
plt.ylim(img_cut_new.shape[0], 0)
mask = img_cut_new < 1.5 * np.median(img_cut_new) # Background / weak stars subtraction
img_cut_new[mask] = 0
segmap = label(img_cut_new > 0)
fig2 = plt.figure(2)
plt.clf()
plt.imshow(segmap)
plt.figure(1)
cent_new = []
for i_object in range(1, np.max(segmap) + 1):
cent_new.append(centroid(img_cut_new * (segmap == i_object)))
# plt.plot(cent_new[-1][1], cent_new[-1][0], '.', lw=2, color='black')
# if i > 0:
# plt.plot(cent[-1][1], cent[-1][0], '.', lw=2, color='black')
cent_new = np.array(cent_new)
if i > 0:
#print 'Applied offset in pixels:', offset
s = min(len(cent_new), len(cent))
if s > 0:
aux_diff = cent[:s] - cent_new[:s]
# Throw away the differences above the threshold
# mask_cent_diff = np.sqrt(np.sum(aux_diff **2, axis=1)) <= max_offset
mask_cent_diff = np.bitwise_and(max_offset <= cent[:, 1][:s], cent[:, 1][:s] <= guider_size['x_px'] - max_offset)
mask_cent_diff *= np.bitwise_and(max_offset <= cent[:, 0][:s], cent[:, 0][:s] <= guider_size['y_px'] - max_offset)
mask_cent_diff *= np.bitwise_and(max_offset <= cent_new[:, 1][:s], cent_new[:, 1][:s] <= guider_size['x_px'] - max_offset)
mask_cent_diff *= np.bitwise_and(max_offset <= cent_new[:, 0][:s], cent_new[:, 0][:s] <= guider_size['y_px'] - max_offset)
cent_diff = np.median(aux_diff[mask_cent_diff], axis=0)
print mask_cent_diff
print 'Centroid differences:', cent_diff
err = np.sqrt(np.sum((cent_diff - offset) ** 2))
print 'Error:', err
color = 'white' if err == 0 else 'red'
plt.text(5, 15, 'Offset: %s px, Calculated: %s px, Error: %3.2f' % (offset, cent_diff, err),
color=color)
if color == 'red': print 'Attn: %i' % i
for i_object in np.arange(s)[mask_cent_diff]:
plt.plot(cent[i_object][1], cent[i_object][0], '.', lw=10, color='green')
plt.plot(cent_new[i_object][1], cent_new[i_object][0], '.', lw=10, color='blue')
else:
plt.text(5, 15, 'Could not calculate offset!', color='red')
cent = np.copy(cent_new)
fig1.savefig('test/img_%04d.png' % i)
img_cut_new = img_cut_new[px_border:-px_border, px_border:-px_border]
fig1 = plt.figure(1);
plt.clf()
plt.imshow(np.log10(img_cut_new), cmap=plt.cm.gray)
plt.xlim(0, img_cut_new.shape[1])
plt.ylim(img_cut_new.shape[0], 0)
mask = img_cut_new < 1.5 * np.median(img_cut_new) # Background / weak stars subtraction
img_cut_new[mask] = 0
segmap = label(img_cut_new > 0)
fig2 = plt.figure(2);
plt.clf()
plt.imshow(segmap)
plt.figure(1)
cent_new = []
for i_object in range(1, np.max(segmap) + 1):
cent_new.append(centroid(img_cut_new * (segmap == i_object)))
# plt.plot(cent_new[-1][1], cent_new[-1][0], '.', lw=2, color='black')
# if i > 0:
# plt.plot(cent[-1][1], cent[-1][0], '.', lw=2, color='black')
cent_new = np.array(cent_new)
if i > 0:
print 'Applied offset in pixels:', offset
s = min(len(cent_new), len(cent))
if s > 0:
aux_diff = cent[:s] - cent_new[:s]
# Throw away the differences above the threshold
# mask_cent_diff = np.sqrt(np.sum(aux_diff **2, axis=1)) <= max_offset
mask_cent_diff = np.bitwise_and(max_offset <= cent[:, 1][:s], cent[:, 1][:s] <= guider_size['x_px'] - max_offset)
mask_cent_diff *= np.bitwise_and(max_offset <= cent[:, 0][:s], cent[:, 0][:s] <= guider_size['y_px'] - max_offset)
mask_cent_diff *= np.bitwise_and(max_offset <= cent_new[:, 1][:s], cent_new[:, 1][:s] <= guider_size['x_px'] - max_offset)
