no accelerated colorspace conversion found from yuv420p to rgb24.

[anishgoel1]

I'm currently running a very simple test case for image capture, but the resulting image is always black with the message that "no accelerated colorspace conversion found from yuv420p to rgb24." and "no frame". I thought this initially was because of my cv2 build, but after trying multiple times over again, I can't get it to fully work. I'm using the most recent versions of cv2 and DJITelloPy. Here is a sample test case that is producing this:

from djitellopy import tello
import cv2

drone = tello.Tello()
drone.connect()

drone.streamon()
img = drone.get_frame_read().frame
img = cv2.resize(img, (340, 260))
cv2.imwrite("picture.png", img)
drone.streamoff()

upon running this, picture.png is entirely black and the messages no accelerated colorspace conversion found from yuv420p to rgb24 and no frame appear._ I'm running this on a Mac M2 air.

[raulperezalejo]

I can confirm the same result with the same code in a Apple M1 Pro, should be relates with not having graphic acceleration?

[cynthia2006]

This is not an issue of OpenCV, but of FFmpeg, more specifically of libswrescale. This is because aside from x86, hardware accelarate conversion of YUV 4:2:0 to RGB colorspace is not implemented.