Supporting OpenCV as an API for cameras

[alifuaziz]

Issues to be solved

1. Closing the processes which the frames are aquired by properly using the multiprocessing
2. Understanding and fixing why the queue is mostly 2 frames even though it should be lonter sometimes
3. Low resolution output from the cameras (color conversion using cv2?)
4. Opening the camera is slow on the windows computer...
5. Settings and getting camera parameters

[alifuaziz]

Code block for testing just the camera API module

from GUI.GUI_main import GUIMain
from GUI.camera_setup_tab import CameraSettingsConfig
from camera_api.opencv import OpenCVCamera
import time

settings = CameraSettingsConfig(
    unique_id="0-opencv", name=None, fps=30, exposure_time=10000, gain=5, pixel_format="Mono8", downsampling_factor=1
)
cam = OpenCVCamera(settings)
if __name__ == "__main__":
    # Your main code here
    cam.begin_capturing(settings)
    print("wait")
    time.sleep(5)
    print(cam.get_available_images())
    print("closing api")
    cam.close_api()

[alifuaziz]

This is code which is able to run a process and end is correctly usingg the before_exit function on closing the process. We probably want this to de-allocate memory using the self.cap.release() code wherever the process it stops

import multiprocessing
import time
import sys
from signal import signal, SIGTERM
import cv2


def before_exit(*args):
    print("Releasing camera and exiting...")
    if "cap" in globals():
        cap.release()
    cv2.destroyAllWindows()
    sys.exit(0)  # don't forget to exit!


def worker():
    global cap
    signal(SIGTERM, before_exit)
    cap = cv2.VideoCapture(0)

    while True:
        ret, frame = cap.read()
        if not ret:
            break
        cv2.imshow("Camera", frame)
        if cv2.waitKey(1) & 0xFF == ord("q"):
            break

    # before_exit()


if __name__ == "__main__":
    proc = multiprocessing.Process(target=worker, args=())
    proc.start()
    time.sleep(3)
    proc.terminate()
    proc.join()