Why cv2.VideoCapture(1) needs four seconds to detect a movement?

Asked Aug 10 '19 at 18:33

Active Aug 10 '19 at 18:43

Viewed 179 times

This code is running, but it takes four seconds (or four pictures) to detect a movement.

The pictures are 640px x 480px.

Lenovo T520, ELP camera - 1920 x 1080 - ELP Svpro 5-50mm, python3.

I've tried the same code with cap = VideoStream(src=1).start() and it's running perfect.

It's a mystery to me

Why is cap = cv2.VideoCapture(1) so slowly?

ps: one second wait time between pictures is intentional

import cv2


#could be 0 for laptop camera, 1 for second camera
cap = cv2.VideoCapture(1)

count = 0
while True:
    _, frame = cap.read()

    cv2.imshow("Frame", frame)

    key = cv2.waitKey(1000)


    print(count)
    count +=1

    #27 is escape key
    if key == 27:
        break

cap.release()
cv2.destroyAllWindows()

edited Aug 10 '19 at 18:43

asked Aug 10 '19 at 18:33

J. Doe278

OpenCV buffers frames, 5 at a time I think, so you will read old data if you pause 1s between frames. – Mark Setchell Aug 10 '19 at 18:42
Have a look here for solutions... https://stackoverflow.com/a/30032945/2836621 – Mark Setchell Aug 10 '19 at 18:46
@Mark your answer makes sense, now I understand, thanks - ok I check the link – J. Doe278 Aug 10 '19 at 18:50
If you develop a solution, post it as an answer for all to see and learn from, accept it as correct and bag the points. Good luck! – Mark Setchell Aug 10 '19 at 18:52

Why cv2.VideoCapture(1) needs four seconds to detect a movement?

0 Answers0