def transform(diff_image, map_patch):
os.environ['PYOPENCL_COMPILER_OUTPUT'] = '1'
h, w = diff_image.shape[::-1]
pad = np.zeros(map_patch.shape)
pad[:diff_image.shape[0], :diff_image.shape[1]] = diff_image
diff_image = pad
sift_ocl1 = sift.SiftPlan(template=diff_image, devicetype="GPU")
sift_ocl2 = sift.SiftPlan(template=map_patch, devicetype="GPU")
kp1 = sift_ocl1.keypoints(diff_image)
kp2 = sift_ocl2.keypoints(map_patch)
sift_m = sift.MatchPlan(devicetype="GPU")
matches = sift_m.match(kp1, kp2)
sa = sift.LinearAlign(diff_image, devicetype="GPU")
transformed_img = sa.align(map_patch)
transformed_img = transformed_img[:w, :h]
theta_rad = abs(matches.angle[0][0]) + abs(matches.angle[0][1])
theta = (theta_rad * 180) / math.pi
return transformed_img, theta
def img_ex(image1, image2):
os.environ['PYOPENCL_COMPILER_OUTPUT'] = '1'
h1, w1 = image1.shape[:2]
h2, w2 = image2.shape[:2]
image1_resized = cv2.resize(image1,
dsize=(int(w1 / 2), int(h1 / 2)),
interpolation=cv2.INTER_CUBIC)
pad = np.zeros(image1_resized.shape)
pad[:h2, :w2] = image2
image2_resized = pad
sift_ocl1 = sift.SiftPlan(template=image1_resized, devicetype="GPU")
sift_ocl2 = sift.SiftPlan(template=image2_resized, devicetype="GPU")
kp1 = sift_ocl1.keypoints(image1_resized)
kp2 = sift_ocl2.keypoints(image2_resized)
sift_m = sift.MatchPlan(devicetype="GPU")
matches = sift_m.match(kp1, kp2)
u = 0
v = 1
while (v != len(matches)):
if v == len(matches) - 1:
u = u + 1
v = u + 1
dx1 = (matches.x[u][0] - matches.x[v][0])**2
dy1 = (matches.y[u][0] - matches.y[v][0])**2
dx2 = (matches.x[u][1] - matches.x[v][1])**2
dy2 = (matches.y[u][1] - matches.x[v][1])**2
d1 = math.sqrt(dx1 + dy1)
d2 = math.sqrt(dx2 + dy2)
v = v + 1
if d1 < d2 + 10 and d1 > d2 - 10:
break
pt1 = 2 * (matches.y[0][0])
y1 = int(pt1 - 1080)
y2 = int(pt1 + 1080)
if (y1 < 0):
y2 = y2 + abs(y1)
y1 = 0
if (y2 > h1):
y = y2 - h1
y2 = h1
y1 = y1 - y
image1 = image1[y1:y2, 0:w1]
return image1, y1
def __createSiftData(self, image, second_image):
"""Generate key points and aligned images from 2 images.
If no keypoints matches, unaligned data are anyway returns.
:rtype: Tuple(numpy.ndarray,numpy.ndarray)
"""
devicetype = "GPU"
# Compute base image
sift_ocl = sift.SiftPlan(template=image, devicetype=devicetype)
keypoints = sift_ocl(image)
# Check image compatibility
second_keypoints = sift_ocl(second_image)
mp = sift.MatchPlan()
match = mp(keypoints, second_keypoints)
_logger.info("Number of Keypoints within image 1: %i" % keypoints.size)
_logger.info(" within image 2: %i" %
second_keypoints.size)
self.__matching_keypoints = (match[:].x[:, 0], match[:].y[:, 0],
match[:].scale[:, 0])
matching_keypoints = match.shape[0]
_logger.info("Matching keypoints: %i" % matching_keypoints)
if matching_keypoints == 0:
return image, second_image
# TODO: Problem here is we have to compute 2 time sift
# The first time to extract matching keypoints, second time
# to extract the aligned image.
# Normalize the second image
sa = sift.LinearAlign(image, devicetype=devicetype)
data1 = image
# TODO: Create a sift issue: if data1 is RGB and data2 intensity
# it returns None, while extracting manually keypoints (above) works
result = sa.align(second_image, return_all=True)
data2 = result["result"]
self.__transformation = self.__toAffineTransformation(result)
return data1, data2
img2 = cv2.rotate(data.images[0], cv2.ROTATE_90_CLOCKWISE)
sift_ocl = sift.SiftPlan(template=img, devicetype="GPU")
keypoints_1 = sift_ocl.keypoints(img1)
keypoints_2 = sift_ocl.keypoints(img2)
figure, ax = plt.subplots(
1,
2,
)
ax[0].imshow(img1, cmap='gray')
ax[0].plot(keypoints_1[:].x, keypoints_1[:].y, ".r")
ax[1].imshow(img2, cmap='gray')
ax[1].plot(keypoints_2[:].x, keypoints_2[:].y, ".r")
mp = sift.MatchPlan()
matching_keypoints = mp(keypoints_1, keypoints_2)
print("Number of Keypoints - image 1 :", keypoints_1.size, " - image 2 : ",
keypoints_2.size, " - Matching keypoints : ",
matching_keypoints.shape[0])
img3 = cv2.hconcat([img1, img2])
fig, ax = plt.subplots()
ax.imshow(img3, cmap='gray')
ax.axis('off')
ax.plot(matching_keypoints[:, 0].x, matching_keypoints[:, 0].y, '.r')
ax.plot(matching_keypoints[:, 1].x + img1.shape[1], matching_keypoints[:, 1].y,
'.r')
for m in matching_keypoints:
ax.arrow(m[0].x, m[0].y, m[1].x - m[0].x + img1.shape[1], m[1].y - m[0].y)
fig.show()