A Real-Time Architecture of 360-Degree Panoramic Video Streaming System
Myungwoo Han, Suk‐Hwan Lee, Soo-Yol Ok · 2019
In this paper, we discuss how to develop a system architecture that produces and transmits 360 degree panoramic video data with multiple cameras in real-time. Our system consists of three stages; panoramic video production using ORB(Oriented FAST and rotated BRIEF)[1] feature point detection, H.264 video compression format and RTSP(Real Time Streaming Protocol) streaming. In particular, 360 degree panoramic video is produced by acceleration of camera calibration, stitching, blending and encoding processes by GPU parallel programming named “CUDA”. We measured the fps(frames per second) of panoramic video streaming that is transmitted to the client and verified that our system has at least 30fps at 4K output resolution. Thus, we could see that our system is suitable for 360 degree panoramic video streaming service in real time.