CODEWITHSUNDEEP
pythonpython-3.xopencv
timonsku
timonsku

Reputation: 1259

OpenCV unable to read video from MJPEG stream

I'm trying to open an mjpeg stream with VideoCapture in OpenCV2 But whenever I try to read a frame I get the following error thrown: [mjpeg @ 0x10f4d20] unable to decode APP fields: Invalid data found when processing input

I can watch the stream without issues in the browser. I also tried the typical suggestion of adding a dummy parameter like ?type=.mjpg but no luck.

This is how I open the stream:

cap = cv2.VideoCapture("http://localhost:8000/camera/mjpeg?type=.mjpg")
while cap.isOpened():
    ret, image = cap.read()
    if not ret:
        break
    cv2.imshow("Result", image)

Upvotes: 1

Views: 7076

Answers (3)

NineStars
NineStars

Reputation: 1

If you are using mjpeg streamer, to access to the stream use

stream = opencv.VideoCapture('http://localhost:8080/?action=stream')

or

stream = opencv.VideoCapture('http://XX.XX.XX.XX:8080/?action=stream')

Upvotes: 0

user4589641
user4589641

Reputation:

I am doing exactly this (Python 3.7) and it is working. I have a Raspberry Pi 4 sending out a stream. On the same network, my MacBook is running the following code.

# Open a URL stream
stream = cv2.VideoCapture('http://192.168.50.1:8080/stream.mjpg')
while ( stream.isOpened() ):
    # Read a frame from the stream
    ret, img = stream.read()
    if ret: # ret == True if stream.read() was successful
        cv2.imshow('Video Stream Monitor', img)

So I'm not sure what you were doing wrong.

Upvotes: 1

Rifat Alptekin Çetin
Rifat Alptekin Çetin

Reputation: 1298

you need to use urllib for reading that

import cv2
import urllib.request
import numpy as np

stream = urllib.request.urlopen('http://localhost:8000/camera/mjpeg?type=.mjpg')
bytes = b''
while True:
    bytes += stream.read(1024)
    a = bytes.find(b'\xff\xd8') #frame starting 
    b = bytes.find(b'\xff\xd9') #frame ending
    if a != -1 and b != -1:
        jpg = bytes[a:b+2]
        bytes = bytes[b+2:]
        img = cv2.imdecode(np.fromstring(jpg, dtype=np.uint8), cv2.CV_LOAD_IMAGE_COLOR)
        cv2.imshow('image', img)
        if cv2.waitKey(1) == 27:
            cv2.destroyAllWindows()
            break

Upvotes: 0

Related Questions