Hybrid CPU-GPU Acceleration of a Multithreaded Image Stitching Algorithm


Tesfay S. W., Demirdag Z. G., Ugurdag H. F., Ates H. F.

7th International Conference on Computer Science and Engineering, UBMK 2022, Diyarbakır, Turkey, 14 - 16 September 2022, pp.468-473 identifier

  • Publication Type: Conference Paper / Full Text
  • Doi Number: 10.1109/ubmk55850.2022.9919473
  • City: Diyarbakır
  • Country: Turkey
  • Page Numbers: pp.468-473
  • Keywords: CPU-GPU, Image stitching, Multithreading, Real-time image processing
  • Istanbul Medipol University Affiliated: Yes

Abstract

Real-time image stitching is critical, especially in un-manned aerial vehicles, and its acceleration has received attention in recent years. This paper describes an image stitching acceleration scheme for heterogeneous (CPU+GPU) devices. Acceleration is attempted with both multithreading and multiprocessing. Most time-critical functions in the algorithm are offloaded on to the GPU. We crafted a 3-buffer ping-pong mechanism for synchro-nization and data transfer among threads/processes in order to maximize CPU utilization. We carried out our experiments on Nvidia Jetson AGX Xavier. Results show that more than 3x acceleration is achieved.