OpenCV Python - 使用RANSAC的findHomography

Question

I'm trying to run findHomography() using RANSAC, but I'm receiving a TypeError. My code works when I don't specify method at all, but even method=0 causes a TypeError.

Code

    T, mask = cv2.findHomography(points_subset[i], points_subset[i+1], False, method=cv2.RANSAC)

Error

Traceback (most recent call last):
  File "stab.py", line 368, in <module>
    stabilize_video(path, video_name)
  File "stab.py", line 331, in stabilize_video
    transforms, points, frame = calc_transformations(cap)
  File "stab.py", line 113, in calc_transformations
    T, mask = cv2.findHomography(points_subset[i], points_subset[i+1], False, method=cv2.RANSAC)
TypeError: Argument given by name ('method') and position (3)

Answer 1

Your error is occurring because you are attempting to specify the method parameter twice; once as a positional argument, and again as a keyword argument. The False argument can be removed to correct your error. If you are trying to use RANSAC to find the homography, the correct call looks like this:

T, mask = cv2.findHomography(points_subset[i], points_subset[i+1], method=cv2.RANSAC)