def rectify_aoi(file1, file2, aoi, z=None, correct_pointing=True,
register_ground=True, debug=False):
"""
Args:
file1, file2 (strings): file paths or urls of two satellite images
aoi (geojson.Polygon): area of interest
z (float, optional): base altitude with respect to WGS84 ellipsoid. If
None, z is retrieved from srtm.
Returns:
rect1, rect2: numpy arrays with the images
S1, S2: transformation matrices from the coordinate system of the original images
disp_min, disp_max: horizontal disparity range
P1, P2: affine rpc approximations of the two images computed during the rectification
"""
# read the RPC coefficients
rpc1 = utils.rpc_from_geotiff(file1)
rpc2 = utils.rpc_from_geotiff(file2)
# get the altitude of the center of the AOI
if z is None:
lon, lat = np.mean(aoi['coordinates'][0][:4], axis=0)
z = srtm4.srtm4(lon, lat)
# compute rectifying affine transforms
S1, S2, w, h, P1, P2 = rectifying_affine_transforms(rpc1, rpc2, aoi, z, register_ground)
# compute SIFT keypoint matches (needed to estimate the disparity range)
q1, q2 = sift_roi(file1, file2, aoi, z)
# correct pointing error with the SIFT keypoint matches (optional)
if correct_pointing:
S1, S2 = pointing_error_correction(S1, S2, q1, q2)
# rectify the crops
rect1 = affine_crop(file1, S1, w, h)
rect2 = affine_crop(file2, S2, w, h)
#print (file1,S1,w,h)
# transform the matches to the domain of the rectified images
q1 = utils.points_apply_homography(S1, q1)
q2 = utils.points_apply_homography(S2, q2)
# disparity range bounds
kpts_disps = (q2 - q1)[:, 0]
disp_min = np.percentile(kpts_disps, 5)
disp_max = np.percentile(kpts_disps, 100 - 5)
if debug: # matches visualisation
import cv2
kp1 = [cv2.KeyPoint(x, y, 1, 0) for x, y in q1]
kp2 = [cv2.KeyPoint(x, y, 1, 0) for x, y in q2]
matches = [[cv2.DMatch(i, i, 0, 0)] for i in range(len(q1))]
plt.figure()
plt.imshow(cv2.drawMatchesKnn(utils.simple_equalization_8bit(rect1), kp1,
utils.simple_equalization_8bit(rect2), kp2,
matches, None, flags=2))
plt.show()
return rect1, rect2, S1, S2, disp_min, disp_max, P1, P2
def affine_crop(input_path, A, w, h):
"""
Apply an affine transform to an image.
Args:
input_path (string): path or url to the input image
A (numpy array): 3x3 array representing an affine transform in
homogeneous coordinates
w, h (ints): width and height of the output image
Return:
numpy array of shape (h, w) containing a subset of the transformed
image. The subset is the rectangle between points 0, 0 and w, h.
"""
# determine the rectangle that we need to read in the input image
output_rectangle = [[0, 0], [w, 0], [w, h], [0, h]]
x, y, w0, h0 = utils.bounding_box2D(utils.points_apply_homography(np.linalg.inv(A),
output_rectangle))
x, y = np.floor((x, y)).astype(int)
w0, h0 = np.ceil((w0, h0)).astype(int)
# crop the needed rectangle in the input image
with rasterio.open(input_path, 'r') as src:
aoi = src.read(indexes=1, window=((y, y + h0), (x, x + w0)), boundless=True)
# compensate the affine transform for the crop
B = A @ utils.matrix_translation(x, y)
# apply the affine transform
out = ndimage.affine_transform(aoi.T, np.linalg.inv(B), output_shape=(w, h)).T
return out
def rectify_aoi(file1, file2, aoi, z=None):
"""
Args:
file1, file2: filename of two satellite images
aoi: area of interest
z (float, optional): base altitude with respect to WGS84 ellipsoid. If
None, z is retrieved from srtm.
Returns:
rect1, rect2: numpy arrays with the images
S1, S2: transformation matrices from the coordinate system of the original images
disp_min, disp_max: horizontal disparity range
P1, P2: affine rpc approximations of the two images computed during the rectification
"""
# read the RPC coefficients
rpc1 = utils.rpc_from_geotiff(file1)
rpc2 = utils.rpc_from_geotiff(file2)
# get the altitude of the center of the AOI
if z is None:
lon, lat = np.mean(aoi['coordinates'][0][:4], axis=0)
z = srtm4.srtm4(lon, lat)
# compute rectifying affine transforms
S1, S2, w, h, P1, P2 = rectifying_affine_transforms(rpc1, rpc2, aoi, z=z)
# compute sift keypoint matches
q1, q2 = sift_roi(file1, file2, aoi, z)
# transform the matches to the domain of the rectified images
q1 = utils.points_apply_homography(S1, q1)
q2 = utils.points_apply_homography(S2, q2)
# pointing correction (y_shift)
y_shift = np.median(q2 - q1, axis=0)[1]
S2 = matrix_translation(0, -y_shift) @ S2
# rectify the crops
rect1 = affine_crop(file1, S1, w, h)
rect2 = affine_crop(file2, S2, w, h)
# disparity range bounds
kpts_disps = (q2 - q1)[:, 0]
disp_min = np.percentile(kpts_disps, 2)
disp_max = np.percentile(kpts_disps, 100 - 2)
return rect1, rect2, S1, S2, disp_min, disp_max, P1, P2
def sift_roi(file1, file2, aoi, z):
"""
Args:
file1, file2 (str): paths or urls to two GeoTIFF images
aoi (geojson.Polygon): area of interest
z (float): base altitude for the aoi
Returns:
two numpy arrays with the coordinates of the matching points in the
original (full-size) image domains
"""
# image crops
crop1, x1, y1 = utils.crop_aoi(file1, aoi, z=z)
crop2, x2, y2 = utils.crop_aoi(file2, aoi, z=z)
# sift keypoint matches
p1, p2 = match_pair(crop1, crop2)
q1 = utils.points_apply_homography(utils.matrix_translation(x1, y1), p1)
q2 = utils.points_apply_homography(utils.matrix_translation(x2, y2), p2)
return q1, q2
def pointing_error_correction(S1, S2, q1, q2):
"""
Correct rectifying similarities for the pointing error.
Args:
S1, S2 (np.array): two 3x3 matrices representing the rectifying similarities
q1, q2 (lists): two lists of matching keypoints
Returns:
two 3x3 matrices representing the corrected rectifying similarities
"""
# transform the matches to the domain of the rectified images
q1 = utils.points_apply_homography(S1, q1)
q2 = utils.points_apply_homography(S2, q2)
# CODE HERE: insert a few lines to correct the vertical shift
y_shift = np.median(q2 - q1, axis=0)[1]
S1 = utils.matrix_translation(-0, +y_shift / 2) @ S1
S2 = utils.matrix_translation(-0, -y_shift / 2) @ S2
return S1, S2
def sift_roi(file1, file2, aoi, z):
"""
Args:
file1, file2: filename of two satellite images
aoi: area of interest
z: base height for the aoi
Returns:
q1, q2: numpy arrays with the coordinates of the matching points in the
original (full-size) image domains
"""
# image crops
crop1, x1, y1 = utils.crop_aoi(file1, aoi, z=z)
crop2, x2, y2 = utils.crop_aoi(file2, aoi, z=z)
# sift keypoint matches
p1, p2 = match_pair(crop1, crop2)
q1 = utils.points_apply_homography(matrix_translation(x1, y1), p1)
q2 = utils.points_apply_homography(matrix_translation(x2, y2), p2)
return q1, q2