首页 | 本学科首页   官方微博 | 高级检索  
     检索      

FPGA+DSP航片并行处理系统
引用本文:雷海军,李德华,薛雷,关景火.FPGA+DSP航片并行处理系统[J].华中科技大学学报(自然科学版),2002,30(11):28-30.
作者姓名:雷海军  李德华  薛雷  关景火
作者单位:华中科技大学图像识别与人工智能研究所,图像信息处理与智能控制教育部重点实验室
摘    要:研究了一种采用FPGA +双DSP的航片高速并行处理系统 ,并用区域分解算法对航片处理任务进行划分与分配 .FPGA实现对航片预处理 .DSP实现航片高层处理 .DSP部分由双TMS32 0C6 2 0 1芯片构成高速运算处理单元 ,峰值处理能力每s可达 3.2× 10 9条指令 .FPGA和DSP具有各自的存储器 .在系统中应用符合数字图像处理特点的区域分解并行算法 ,这样使在空间域串行图像处理算法得到并行化 ,从而合理地对任务进行划分与分配 ,同时保证各DSP处理机负载平衡 .该方法适合多种图像处理算法 ,实现简单 ,大大减少了开发的工作量 .经试验表明 ,该实时航片处理系统具有高效、简单、可靠的特点 .

关 键 词:DSP  航片并行处理系统  系统结构  FPGA  数字信号处理器  区域分解算法  图像处理  目标识别
文章编号:1671-4512(2002)11-0028-03
修稿时间:2002年5月27日

Aerial image parallel processing system based on FPGA+DSP
Lei Haijun Li Dehua Xue Lei Guan Jinghuo Doctoral Candidate,Institute for Pattern Recognition and Artificial Intelligence,Huazhong Univ . of Sci. & Tech.,Wuhan ,China..Aerial image parallel processing system based on FPGA+DSP[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,2002,30(11):28-30.
Authors:Lei Haijun Li Dehua Xue Lei Guan Jinghuo Doctoral Candidate  Institute for Pattern Recognition and Artificial Intelligence  Huazhong Univ of Sci & Tech  Wuhan  China
Institution:Lei Haijun Li Dehua Xue Lei Guan Jinghuo Doctoral Candidate,Institute for Pattern Recognition and Artificial Intelligence,Huazhong Univ . of Sci. & Tech.,Wuhan 430074,China.
Abstract:FPGA+DSP is an important model in high speed aerial image parallel system, and the domain decomposition method is applied in partition and allocation of task of the system. The preprocessing algorithms of aerial image are implemented by FPGA and the high level algorithms are performed in DSP. On the basis of the fundamental characters of digital image processing domain decomposition method was used. The method can parallel serial digital image processing algorithms which process digital images on spatial domain and implement partition and allocation of processing task of aerial image. Moreover, it ensures DSP load balance in the course of processing aerial image and applies to many image algorithms with no difficulty. Experimental results show that our real time system is effective, simple and steady.
Keywords:aerial image  system architecture  real time system  FPGA  DSP  domain decomposition method
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号