PiCamera和CV2解码总是返回None

Question

I have a problem using CV2 on Raspberry PI with the PiCamera. This is my code:

camera = PiCamera()
camera.resolution = ( 320, 240 )
camera.framerate = 60
rawCapture = PiRGBArray( camera, size=( 320, 240 ) )
time.sleep(1)
start = time.time()

for frame in camera.capture_continuous(rawCapture, format="bgr", use_video_port=True):


    # and occupied/unoccupied text
    images = frame.array

    # show the frame
    cv2.imshow("Frame", images)


    cv_img = images.astype(np.uint8)

    gray  = cv2.imdecode(np.fromstring(cv_img, dtype=np.uint8), cv2.IMREAD_GRAYSCALE)
    image = cv2.imdecode(np.fromstring(cv_img, dtype=np.uint8), cv2.IMREAD_UNCHANGED)

    print("IMAGE: ",image)
    print("GRAY: ",gray)

image and gray are always None while images and cv_image contains the right frame. I can't really understand why it won't work, Could you please help me?

Thank you in advance.

Answer 1

I've found picamera.array to be a reliable way to bridge between picamara and cv2 . The docs have some examples that aren't far off from yours.

Once you're squarely in cv2 territory,

gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)

is the typical way to do the grayscale conversion. Doing it by way of imdecode is excessive.