You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have same issue, and I recognized that the matrix H(returned from cv2.findHomography) was None.
I found that cv2.findHomography can return None if the matrix H cannot be calculated but I could not find any solution for that..
Is anybody know how to solve the problem?
This may happen if the background is featureless, like a flat wall. For homography, it needs to find sufficient keypoints via AKAZE features, match them reliably and calculate homography (needs at least 8 points), if not it will break.
There is no problem in running the alignment part of the demo.What is the reason for the following error when running my own picture?
python test_pre_process.py -i sample_data/my_photo/
Traceback (most recent call last):
File "test_pre_process.py", line 108, in
back_align = alignImages(back, image,mask)
File "test_pre_process.py", line 42, in alignImages
im1Reg = cv2.warpPerspective(im1, h, (width, height))
cv2.error: OpenCV(3.4.5) /io/opencv/modules/imgproc/src/imgwarp.cpp:2927: error: (-215:Assertion failed) (M0.type() == CV_32F || M0.type() == CV_64F) && M0.rows == 3 && M0.cols == 3 in function 'warpPerspective'
The text was updated successfully, but these errors were encountered: