OpenCV 4.4.0.44 macOS Catalina - Hang/Frozen

[JapinderSandhu]

System information (version)

• OpenCV => 4.4.0.44
• numpy => 1.19.2
• Operating System / Platform => macOS Catalina Version 10.15.6
• Code Editor => Visual Studio Code 2020
• Compiler => Anaconda 4.7.12 & 4.7.10

Detailed description

I am trying to convert a video frame to gray using cvtColor. This code works perfectly fine in macOS Mojave, MacOS High Sierra but Breaks in macOS Catalina. Other than these I have not tested other macOS. This code is not meant for Windows so I have not tested it on that.

The video frame is loading successfully, cap.read() is returning a frame with boolean True, so the video file path is correct and confirmed.

I tried #18079

openCV 4.3.0.36 as mentioned in #18079

The code stops at "BEFORE CVT GRAY" and never prints "AFTER CVT GRAY"

Steps to reproduce

code snippet

        ##############################################################
        #  Step 2: Read the first frame and convert it to grayscale
        ##############################################################

        cap = cv2.VideoCapture(input["mediaPath"] )

        # set frame to startFrame
        cap.set(cv2.CAP_PROP_POS_FRAMES, 0)

        # Read first frame
        ret, prev = cap.read()

        print("FRAME READ",ret)

        if not ret:

            print("Error cant read frames")

        print("IMAGE TEST", prev)

        print("BEFORE CVT GRAY")
        # Convert frame to grayscale
        prev_gray = cv2.cvtColor(prev, cv2.COLOR_BGR2GRAY)
        print("AFTER CVT GRAY")

Issue submission checklist

• [ ✅] I report the issue, it's not a question
• [ ✅] I checked the problem with documentation, FAQ, open issues,
  answers.opencv.org, Stack Overflow, etc and have not found solution
• [:white_check_mark: ] I updated to latest OpenCV version and the issue is still there
• [:white_check_mark: ] There is reproducer code and related data files: videos, images, onnx, etc

[alalek]

Provide complete minimal reproducer. Exclude unnecessary parts if they works (like VideoCapture, use np.zeros() instead).

code stops

Please attach stack trace.

[JapinderSandhu]

SOLVED: added line in server for multiprocessing

this code snippet was in a larger multiprocessing function with a server so thats why I was unable to include it all for minimal reproducer. But I found the solution here.

Turns out I needed to add the line

multiprocessing.set_start_method('spawn')

#5150