A parallel image restoration and super-resolution processing system based on the PPF and Cluster architecture

Dong-Dong MA; Jin-zong LI; Bing ZHU; Li-sheng MU

doi:null

您当前的位置：

首页 >

文章列表页 >

A parallel image restoration and super-resolution processing system based on the PPF and Cluster architecture

Article | 更新时间：2020-08-13

- A parallel image restoration and super-resolution processing system based on the PPF and Cluster architecture
- Optics and Precision Engineering Vol. 17, Issue 5, Pages: 1149-1160(2009)
- 作者机构：
  
  哈尔滨工业大学
- 作者简介：
- 基金信息：
- DOI：
  CLC：
- Received：18 July 2008，
  
  Revised：17 September 2008，
  
  Published Online：25 May 2009，
  
  Published：25 May 2009
- 稿件说明：
移动端阅览
马冬冬,李金宗,朱兵,穆立胜. 基于PPF和机群结构的并行图象复原与超分辨处理系统[J]. 光学精密工程, 2009,17(5):1149-1160

Dong-Dong MA, Jin-zong LI, Bing ZHU, et al. A parallel image restoration and super-resolution processing system based on the PPF and Cluster architecture[J]. Optics and precision engineering, 2009, 17(5): 1149-1160.
马冬冬,李金宗,朱兵,穆立胜. 基于PPF和机群结构的并行图象复原与超分辨处理系统[J]. 光学精密工程, 2009,17(5):1149-1160 DOI：

Dong-Dong MA, Jin-zong LI, Bing ZHU, et al. A parallel image restoration and super-resolution processing system based on the PPF and Cluster architecture[J]. Optics and precision engineering, 2009, 17(5): 1149-1160. DOI：

摘要

为实现海量大尺度图象的复原与超分辨算法的实时处理，完成了相应并行系统的设计，对该系统的技术方案、工作原理、关键算法、硬件结构和并行技术等进行了研究。首先，根据技术方案介绍了系统的工作原理及解模糊、去噪和超分辨等关键算法。接着，对DSP和机群两种体系结构进行了分析和比较，结果表明机群更适合于大规模并行处理，并给出其设备选型的原则。然后，提出了基于PPF结构的并行算法模型和基于MPI＋OpenMP混合结构的多层次并行与优化技术。最后，对算法在DSP和计算机系统上的处理效果和处理速度进行了实验和分析，给出了系统规模和性能的预测，确定了关键参数即处理器数目的选择依据。实验结果表明，该系统可将处理时间由2700s降低到29.39s，同时使处理后图象的清晰度、对比度和分辨率显著提高，满足应用需求并具有一定的通用性。

Abstract

A parallel processing system was developed to realize real-time restoration and super-resolution processing of huge amount of large size images after investigating its technical scheme

working principle

key algorithms

hardware architecture

parallel technology and et al. First

the system’s working principle and key algorithms of deblur

denoise

super-resolution were presented based on the technical scheme. After analyzing and comparing the two architectures of DSP and Cluster

employ principle of the Cluster’s equipments was given. It is concluded that Cluster is more suited for large scale parallel processing. Then a parallel algorithm model based on the PPF architecture and multi-layer parallel optimization technology based on the hybrid architecture of MPI+OpenMP were introduced. Finally

experiments and analyses were made on processing effect and speed of DSP and computer

following by scale and performance prediction of system. How to choice the number of processes was also discussed. Experimental results show that the parallel system can reduce run time from 2700s to 29.39s

making images improved apparently in sharpness

contrast and resolution. It can satisfy our application and have a universal property。

关键词

Keywords

references

Views

1311

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

DC position close-loop system of spectrometer by feed-forward correction

Study on Palm-print Images Sampling Equimpment

Algorithms of Particle Swarm Optimizer Tracking Based on DSP Parallel System

Study on DSP-based Image Motion Speed Calculation of the Space Camera

Research on Self-testing Method of Spatial Camera Controller Based on Mixed Language Programming

Related Author

WANG Zhi-hong

LIU Jie

LIU Ze-kun

赵继忠

苑玮琦

许廷发（理工）

孔德柱

Related Institution

吉林大学仪器科学与电气工程学院

沈阳工业大学视觉检测研究所

北京理工大学

中科院长春光学精密机械与物理研究所

中国科学院长春光学精密机械与物理研究所

AI问答

Address：No.3888 Dong Nanhu Road, Changchun, Jilin, China Postal code：130033
Tel：0431-86176855 Email：gxjmgc@ciomp.ac.cn
Technical support is provided by Beijing Founder electronics co., LTD 吉ICP备11002662号-17 京公网安备11010802024621
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰