Share Email Print
cover

Proceedings Paper

A distributed multiprocessor system designed for real-time image processing
Author(s): Zhiyi Yin; Wei Heng
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

In real-time image processing, a large amount of data is needed to be processed at a very high speed. Considering the problems faced in real-time image processing, a distributed multiprocessor system is proposed in this paper. In the design of the distributed multiprocessor system, processing tasks are allocated to various processes, which are bound to different CPUs. Several designs are discussed, and making full use of every process is very important to system's excellent performance. Furthermore, the problems of realization fasten on the inter-process communication, the synchronization, and the stability. System analysis and performance tests both show that the distributed multiprocessor system is able to improve system's performance variously, including the delay, the throughput rate, the stability, the scalability. And the system can be expanded easy at aspects of software and hardware. In a word, the distributed multiprocessor system designed for real-time image processing, based on distributed algorithms, not only improves system's performance variously, but also costs low and expands easy.

Paper Details

Date Published: 25 August 2009
PDF: 6 pages
Proc. SPIE 7375, ICEM 2008: International Conference on Experimental Mechanics 2008, 73752M (25 August 2009); doi: 10.1117/12.839223
Show Author Affiliations
Zhiyi Yin, Southeast Univ. (China)
Wei Heng, Southeast Univ. (China)


Published in SPIE Proceedings Vol. 7375:
ICEM 2008: International Conference on Experimental Mechanics 2008
Xiaoyuan He; Huimin Xie; YiLan Kang, Editor(s)

© SPIE. Terms of Use
Back to Top