I am trying to initialize some blank frames of video and write it to a file using opencv and python in ubuntu linux. I am planning to later append in smaller video segments over the blank canvas. The typical examples only show reading a frame from some input source but I am trying to generate my own frame. Here is my code so far:
import numpy as np
import cv2
xsp=640
ysp=480
chan=3
# Define the codec and create VideoWriter object
fourcc = cv2.VideoWriter_fourcc(*'XVID')
out = cv2.VideoWriter('output.avi',fourcc, 20.0, (xsp,ysp))
numframes = 20
videoout = []
for i in range(1,numframes): # init frames
frame = np.zeros((xsp,ysp,chan),np.uint8)
videoout.append(frame)
for frame in videoout:
out.write(frame)
out.release()
Using the code above, the video is not playable. Opening it within a video player gives an error about not being able to demultiplex content.
However, if I initialize the video frame by opening a video file and taking an existing video frame + put a black rectangle over it I am able to get valid video.
def initVideoFrame(videofile):
cap = cv2.VideoCapture(videofile)
ret, frame = cap.read()
ysp,xsp,chan = frame.shape
cv2.rectangle( frame, ( 0,0 ), ( xsp, ysp), ( 0,0,0 ), -1, 8 )
cap.release()
return frame
frame = initVideoFrame('s.mp4')
ysp,xsp,chan = frame.shape
# Define the codec and create VideoWriter object
fourcc = cv2.VideoWriter_fourcc(*'XVID')
out = cv2.VideoWriter('output.avi',fourcc, 20.0, (xsp,ysp))
numframes = 100
videoout = []
for i in range(1,numframes): # init frames
frame = initVideoFrame('s.mp4')
videoout.append(frame)
for frame in videoout:
out.write(frame)
out.release()
It seems that the frames made of array of uint8 zeros is not enough to write a valid video. I am still missing something.