计算机工程与应用 ›› 2009, Vol. 45 ›› Issue (11): 183-185.DOI: 10.3778/j.issn.1002-8331.2009.11.055

• 图形、图像、模式识别 • 上一篇    下一篇

CUDA架构下的快速图像去噪

李 军,李艳辉,陈双平   

  1. 暨南大学 珠海学院 计算机系,广东 珠海 519070
  • 收稿日期:2008-10-09 修回日期:2009-01-12 出版日期:2009-04-11 发布日期:2009-04-11
  • 通讯作者: 李 军

Fast image denoising with CUDA

LI Jun,LI Yan-hui,CHEN Shuang-ping   

  1. Department of Computer,School of Zhuhai,Jinan University,Zhuhai,Guangdong 519070,China
  • Received:2008-10-09 Revised:2009-01-12 Online:2009-04-11 Published:2009-04-11
  • Contact: LI Jun

摘要: 图像处理通常需要较大的计算量,其中图像去噪是经常使用的一种预处理算法,研究其快速算法具有重要意义。图形处理器具有强大的并行计算能力,但大部分时间处于闲置状态。统一计算设备架构提供了一种简单易用的开发环境,可利用图形处理器进行通用计算。提出了基于统一计算设备架构的快速图像去噪算法,可以利用GPU的计算能力,加快去噪过程,显著地减少计算时间。

关键词: 图形处理器, 图像去噪, 统一计算设备架构, 并行数据处理

Abstract: The huge computation load is necessary commonly for the image processing,the research on the fast algorithm of image denoising is very important for it is often used in the field.GPU(Graphic Processing Unit) is powerful for Data-Parallel Processing but it is idle for the most of time.CUDA(Compute Unified Device Architecture) is a simple and easy to use for the general purpose computation.A fast algorithm of image denoising is proposed based on the CUDA,it uses the power of GPU to accelerate the process of image denoising and reduce the time significantly.

Key words: Graphic Processing Unit(GPU), image denoising, Compute Unified Device Architecture(CUDA), data-parallel Processing