0

Essentially, the issue is this; I have 8 security cameras all streaming via RTSP with very high unmodifiable resolutions (specifically, 2560x1440). I want to place a small tag in the corner of the each video stream to display a string of characters with a white background, then save the file locally.

Currently, I am using Python and OpenCV to process the frame, receive the string to apply, modify the frame, then save it to a file. Thing is, the file is already encoded in H264, and I have to retranscode it into H264 and then save it as a file. On a duel core modern Intel i5 (can't change this unfortunately), transcoding eight 2560x1440 video at 15 FPS simultaneously, you can see how this falls apart. I am looking to optimize this process by any means necessary. I also understand this is a very, very tall order. However this is not my decision.

I have created a graphic to help show what I am trying to accomplish.

I would be open to trying something with C++, however I am not experienced with it at all, nor can I find any such libraries for receiving RTSP streams and modifying frames from them.

Any assistance or resources is greatly appreciated.

Julian Nicolai
  • 11
  • 3

1 Answers1

0
  1. You could store your string as bitmap in a separate stream with the correct timing information and do the composite during playback.
  2. You could make the string a subtitle. MP4 and H.264 supports a variety of formats. Then your player just needs to support standard subtitles.
Markus Schumann
  • 7,636
  • 1
  • 21
  • 27