Real-time video image processing through GPUs and CUDA and its future implementation in real problems in a Smart City

  • Alberto Hernández Aguilar Universidad Autónoma del Estado de Morelos
  • Juan Carlos Bonilla-Robles Facultad de Contaduría, Administración e Informática, Universidad Autónoma del Estado de Morelos
  • José C. Zavala Díaz Facultad de Contaduría, Administración e Informática, Universidad Autónoma del Estado de Morelos
  • Alberto Ochoa Maestría en Cómputo Aplicado, Universidad Autónoma de Ciudad Juárez

Abstract

Graphic and coprocessor cards are a fundamental part of digital image processing, since they have supported the CPU in the execution of graphics applications such as 3D video games, computer animation, digital image processing, among other processes that require of a greater number of complex operations for its execution. In this research, we discuss a technological platform in which graphics processing units GPUs, webcams and video, CUDA and free software are used for the treatment of real-time images, for this purpose we describe test software for the application of 3 types of treatment: 3D effect, skin detection and edge detection (Sobel). The results indicate that the higher the complexity level of the effect applied to the image, the higher the percentage of use of the GPUs used: 3D effect (81% of use), edge detection (31% of use) and skin detection (22 %), these values improve significantly this process.

Published
2019-03-12
How to Cite
Hernández Aguilar, A., Bonilla-Robles, J. C., Zavala Díaz, J. C., & Ochoa, A. (2019). Real-time video image processing through GPUs and CUDA and its future implementation in real problems in a Smart City. International Journal of Combinatorial Optimization Problems and Informatics, 10(3), 33-49. Retrieved from https://ijcopi.org/index.php/ojs/article/view/127
Section
Articles