我想在OSX下從OpenCV的python包裝創建視頻。我正在使用python 2.7.1，opencv 2.3.1a以及來自那個opencv版本的willowgarage的python包裝。我有：opencv VideoWriter在OSX下產生沒有輸出
import cv,cv2
w = cv2.VideoWriter('foo.avi', cv.FOURCC('M','J','P','G'), 25, (100,100))
for i in range(100):
w.write(np.ones((100,100,3), np.uint8))
OpenCV的說
WARNING: Could not create empty movie file container.
Didn't successfully update movie file.
... [ 100 repetitions]
我不知道下一個嘗試
在這個話題上有很多過時和不正確的在線指南 - 我想我幾乎每一個都嘗試過。望着基於QTKit代碼實現VideoWriter在Mac OSX後，我終於可以用下面的代碼來獲得VideoWriter到輸出有效的視頻文件：
fps = 15
capSize = (1028,720) # this is the size of my source video
fourcc = cv2.cv.CV_FOURCC('m', 'p', '4', 'v') # note the lower case
self.vout = cv2.VideoWriter()
success = self.vout.open('output.mov',fourcc,fps,capSize,True)
要寫入圖像幀（注意，imgFrame必須尺寸均相同翻船高於或更新會失敗）：
self.vout.write(imgFrame)
完成後一定要：
vout.release()
self.vout = None
這對我的作品在Mac OS X 10.8.5（山獅）：沒有關於其他平臺的保證。我希望這段代碼可以節省別人幾小時的實驗經驗！
這不是好主意
您必須創建一個VideoWriter結構和使用方法WriteFrame ：
import cv
cv.NamedWindow('cap', 1)
w = cv.CreateVideoWriter('test.avi',cv.CV_FOURCC('X','V','I','D'),25,(640,480))
cap = cv.CaptureFromCAM(0)
while(1):
img = cv.QueryFrame(cap)
cv.ShowImage('cap', img)
cv.WriteFrame(w, img)
if cv.WaitKey(1) == 27:
break
cv.DestroyWindow('cap')
這個答案現在看起來已經過時了。 – JohnAllen
您的ffmpeg版本可能已過期。我遇到了與ffmpeg 0.8類似的問題，並且在升級到ffmpeg 0.11後它工作正常。
這裏的@ ToddStellanova的回答爲我工作的變化：
def write_video(image_dir):
'''
Writes a video from a set of images in `image_dir`
'''
target = join('data/labelled-videos',
basename(image_dir) + '.mp4v')
codec = cv2.cv.CV_FOURCC('m', 'p', '4', 'v')
size = (256, 256)
v = cv2.VideoWriter(target, codec, 16, size)
for image_name in listdir(image_dir):
image_filename = join(image_dir, image_name)
arr = np.asarray(Image.open(image_filename))
assert arr.shape[:2] == size
v.write(arr)
後嘗試不同的選擇，我發現，我用不適合在VideoWriter指定大小frame.size：所以將其設置爲默認的我的iMac 1280x720使事情工作！
import numpy as np
import cv2
cap = cv2.VideoCapture(0)
# Define the codec and create VideoWriter object
fourcc = cv2.VideoWriter_fourcc('m', 'p', '4', 'v')
out = cv2.VideoWriter()
succes = out.open('output.mp4v',fourcc, 15.0, (1280,720),True)
while(cap.isOpened()):
ret, frame = cap.read()
if ret==True:
frame = cv2.flip(frame,0)
# write the flipped frame
out.write(frame)
cv2.imshow('frame',frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
else:
break
# Release everything if job is finished
cap.release()
out.release()
cv2.destroyAllWindows()
這是「大小」問題。
import cv2
import time
filename = time.strftime("%m-%d-%H-%M-%S") + '.avi'
fps = 16
cap = cv2.VideoCapture(0)
#in this way it always works, because your get the right "size"
size = (int(cap.get(cv2.cv.CV_CAP_PROP_FRAME_WIDTH)),
int(cap.get(cv2.cv.CV_CAP_PROP_FRAME_HEIGHT)))
fourcc = cv2.cv.FOURCC('8', 'B', 'P', 'S') #works, large
out = cv2.VideoWriter(filename, fourcc, fps, size, True)
#in this way, you must set the "size" to your size,
#because you don't know the default "size" is right or not
#cap.set(3, 640)
#cap.set(4, 480)
#out = cv2.VideoWriter(filename, fourcc, fps, (640, 480), True)
while(cap.isOpened()):
ret, frame = cap.read()
if ret == True:
out.write(frame)
cv2.imshow('frame', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break;
else:
print 'Error...'
break;
cap.release()
out.release()
cv2.destroyAllWindows()
由於@TeddStellanova
這裏的XCode C++版本：
int fWidth = capture.get(CV_CAP_PROP_FRAME_WIDTH);
int fHeight = capture.get(CV_CAP_PROP_FRAME_HEIGHT);
string vid_path = path + "frames.mov";
VideoWriter vid(vid_path,CV_FOURCC('m','p','4','v'),30,Size(fWidth,fHeight),true);
這裏是與工作的版本：
安裝了brew install opencv。
#!/usr/bin/env python3
import cv2
def main():
vc = cv2.VideoCapture()
if not vc.open('input.mp4'):
print('failed to open video capture')
return
fourcc = cv2.VideoWriter_fourcc(*'mp4v')
# Match source video features.
fps = vc.get(cv2.CAP_PROP_FPS)
size = (
int(vc.get(cv2.CAP_PROP_FRAME_WIDTH)),
int(vc.get(cv2.CAP_PROP_FRAME_HEIGHT)),
)
vw = cv2.VideoWriter()
if not vw.open('output.mp4', fourcc, fps, size):
print('failed to open video writer')
return
while True:
ok, frame = vc.read()
if not ok:
break
# Flip upside down.
frame = cv2.flip(frame, 0)
# Write processed frame.
vw.write(frame)
cv2.imshow('frame', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
vc.release()
vw.release()
cv2.destroyAllWindows()
if __name__ == "__main__":
main()
我可以確認這段代碼適用於小牛，Python 2.7，openCV2 –
在到達此處之前嘗試了至少五個其他的答案。這是唯一一個在小牛/ Python 2.7和OpenCV 2.4.6.1上爲我工作的人。 –
與其他評論者的情況相同。我發現決定性的因素是使用小寫的''m'，'p'，'4'，'v'。許多其他示例顯示用於'fourcc'的大寫字母，我不知道它是區分大小寫的。 –