Can't set frame width and height with [OpenCV] cv2.VideoCapture.set()

24,240

Solution 1

The idea is to resize the frame without having to worry about setting the default frame size. Instead of using cv2.VideoCapture().set(), you can use cv2.resize() to resize the original 1920x1080 frame into 320x180. But this method does not maintain aspect ratio. If you wanted to maintain aspect ratio, you can use the imutils library. The imutils.resize() function resizes the frame and maintains aspect ratio. Change the width parameter to your desired resolution

import cv2
import imutils

cap = cv2.VideoCapture(0)

while(cap.isOpened()):
    ret, frame = cap.read()

    frame = imutils.resize(frame, width=320)
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

    cv2.imshow('frame', gray)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()

Solution 2

Finally I have found that solved this problem

When you try to set random resolution, opencv sets nearest resolution if that resolution is not available.

You can run below command to find out all the available resolutions of your webcam.

uvcdynctrl -f

And set those resolutions only. (my webcam resolutions: 1280x720, 640x480, 640x360)

vid.set(cv2.CAP_PROP_FRAME_WIDTH, 640)
vid.set(cv2.CAP_PROP_FRAME_HEIGHT, 480)

Reference: https://techoverflow.net/2018/12/18/how-to-set-cv2-videocapture-image-size/

Share:
24,240
blinkThrice
Author by

blinkThrice

Updated on May 28, 2020

Comments

  • blinkThrice
    blinkThrice almost 4 years

    I am trying to reduce the frame/window size of my video capture to 320x180 but I can't seem to do it. I am using a Windows Kinect for Xbox One and its connected to my pc using an adapter.

    I have tried setting the cv2.CAP_PROP_FRAME_WIDTH to 320 and cv2.CAP_PROP_FRAME_HEIGHT to 180 but once I try and get the values it returns 1920 and 1080. I have also tried installing and reinstalling the Kinect SDK and drivers.

    import cv2
    import numpy as np
    
    vid = cv2.VideoCapture(0)
    
    vid.set(cv2.CAP_PROP_FRAME_WIDTH, 320)
    vid.set(cv2.CAP_PROP_FRAME_HEIGHT, 180)
    vid.set(cv2.CAP_PROP_FPS, 25)
    
    print(vid.get(cv2.CAP_PROP_FPS))
    print(vid.get(cv2.CAP_PROP_FRAME_WIDTH))
    
    while True:
        ret, frame = vid.read()
        gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    
        # Display the resulting frame
        cv2.imshow('frame', gray)
        if cv2.waitKey(1) & 0xFF == ord('q'):
            break
    
    vid.release()
    cv2.destroyAllWindows()
    

    print output

    I would like help in knowing where the problem stems from and hopefully get a fix.

  • pentanol
    pentanol over 2 years
    This does not work
  • ati ince
    ati ince almost 2 years
    nice solution. I am transferring cv image<-> qtlabel pixmap and it solved my need.