A = images[0]
M_0A = np.eye(3)
for i, img in enumerate(images[1:]):
print "Registering %d -> %d" % (i, i + 1)
B = images[i + 1]
M_0B, S = register.dense_MI(B, A, levels=options.levels + 1,
std=3, win_size=9,
translation_only=options.translation,
fixed_scale=options.fixed_scale)
print "Mutual information: ", S
if S > 1.5:
print "Warning: registration (%d -> %d) probably failed." % (i, i + 1)
H = '%03d.%03d.H' % (i, i + 1)
imsave(os.path.join(H_dir, H + '.png'),
register.stack.with_transform([A, B],
[np.eye(3), np.linalg.inv(M_0B)],
oshape=A.shape),
)
M_A0 = np.linalg.inv(M_0A)
M_0A = M_0B
# Transformation relative to previous frame
#
# H_0B = H_AB * H_0A
# => H_AB = H_0B * H_0A.I
#
M_AB = np.dot(M_0B, M_A0)
np.savetxt(os.path.join(H_dir, H), M_AB, delimiter=' ')
for i, img in enumerate(images[1:]):
print "Registering %d -> %d" % (i, i + 1)
B = images[i + 1]
M_0B, S = register.dense_MI(B,
A,
levels=options.levels + 1,
std=3,
win_size=9,
translation_only=options.translation,
fixed_scale=options.fixed_scale)
print "Mutual information: ", S
if S > 1.5:
print "Warning: registration (%d -> %d) probably failed." % (i, i + 1)
H = '%03d.%03d.H' % (i, i + 1)
imsave(
os.path.join(H_dir, H + '.png'),
register.stack.with_transform(
[A, B], [np.eye(3), np.linalg.inv(M_0B)], oshape=A.shape),
)
M_A0 = np.linalg.inv(M_0A)
M_0A = M_0B
# Transformation relative to previous frame
#
# H_0B = H_AB * H_0A
# => H_AB = H_0B * H_0A.I
#
M_AB = np.dot(M_0B, M_A0)
np.savetxt(os.path.join(H_dir, H), M_AB, delimiter=' ')
if options.previous_result:
plt.figure()
plt.plot(err)
plt.xlabel('Iteration')
plt.ylabel('Error Norm')
plt.title('SR Convergence')
if options.photo_adjust:
a, b = photometric_adjust(out, avg)
out *= a
out += b
out = np.clip(out, 0, 255)
imsave('/tmp/avg.png', avg)
print 'Average image saved as /tmp/avg.png'
imsave('/tmp/out.png', out)
print 'Resulting image saved as /tmp/out.png'
plt.figure()
plt.subplot(3, 1, 1)
plt.imshow(ic[0], interpolation='nearest', cmap=plt.cm.gray)
plt.subplot(3, 1, 2)
plt.imshow(avg, interpolation='lanczos', cmap=plt.cm.gray)
plt.subplot(3, 1, 3)
plt.imshow(out, interpolation='lanczos', cmap=plt.cm.gray)
plt.show()
if options.previous_result:
plt.figure()
plt.plot(err)
plt.xlabel('Iteration')
plt.ylabel('Error Norm')
plt.title('SR Convergence')
if options.photo_adjust:
a, b = photometric_adjust(out, avg)
out *= a
out += b
out = np.clip(out, 0, 255)
imsave('/tmp/avg.png', avg)
print 'Average image saved as /tmp/avg.png'
imsave('/tmp/out.png', out)
print 'Resulting image saved as /tmp/out.png'
plt.figure()
plt.subplot(3, 1, 1)
plt.imshow(ic[0], interpolation='nearest', cmap=plt.cm.gray)
plt.subplot(3, 1, 2)
plt.imshow(avg, interpolation='lanczos', cmap=plt.cm.gray)
plt.subplot(3, 1, 3)
plt.imshow(out, interpolation='lanczos', cmap=plt.cm.gray)
plt.show()
