Creating a video from images using OpenCV with Python

Question

My code:

video = cv2.VideoWriter(name, 0, 1, size)

for image in frames:
    video.write(image)

video.release()

name is a string describing the filename, size is (360, 640) and frames is a list containing all the images (ndarrays with all the color values). When I try to run this code, it returns

TypeError: Expected Ptr<cv::UMat> for argument 'image'

What does that mean and how do I fix this error?

Edit: My code is now:

fourcc = cv2.VideoWriter_fourcc(*'XVID')
video = cv2.VideoWriter('mosaic.avi', fourcc, 30, size)

for image in frames:
    video.write(np.float32(image))

video.release()

It does not return any errors, however the video is corrupted.

Answer 1

UMat is a part of the Transparent API (TAPI) than help to write one code for the CPU and OpenCL implementations. You can correct this error by following methods:

video = cv2.VideoWriter(name, 0, 1, size)

for image in frames:
    video.write(cv2.UMat(image))

video.release()

Or

Import numpy as np
video = cv2.VideoWriter(name, 0, 1, size)

for image in frames:
    image = np.array(image)
    video.write(image)

video.release()