I've been trying to optimize performance of a camera. I'm only streaming over RTMP once a particular object is detected from Yolo11. The only working way to do this - from what I can tell so far - is to have two Camera Objects open.
ex:
detector = nn.YOLO11(model="/root/models/yolo11n.mud", dual_buff=True)
cam_yolo = camera.Camera(width,height,detector.input_format())
cam = camera.Camera(width,height,image.Format.FMT_YVU420SP)
Using this method removes a frame from the buffer whenever you call ".read()". Ideally, I'd have just the "cam" camera object and simply convert the returned image. This wouldn't result in a dropped frame the RTMP stream.
Is there a way to convert from the FMT_YVU420SP to the FMT_RGB888 color spaces, or vice versa?
Thanks!