首页 | 本学科首页   官方微博 | 高级检索  
     

基于硬件加速的高精度实时图像处理系统设计
引用本文:王哲,张晓林,刘荣科,王冬,赵广胜.基于硬件加速的高精度实时图像处理系统设计[J].计算机测量与控制,2007,15(3):387-389.
作者姓名:王哲  张晓林  刘荣科  王冬  赵广胜
作者单位:北京航空航天大学电子信息工程学院,北京100083
摘    要:高精度实时图像处理在众多领域中具有广泛的应用前景;在分析高精度实时图像处理要求的基础卜,设汁实现了基于硬件加速的高精度实时图像处理系统;系统采用浮点嵌入式硬件协处理单元对高精度图像处理进行加速;协处理单元和主机之间的高速数据通路和控制通路,便于主机与协处理单元的数据传输及对协处理单元进行软封装;并在此基础上通过软件设计实现了图像的流水线并行处理,可明显提高系统处理速度;通过实验证明,对于文中采用的算法,相比主机本系统可将图像处理速度提高30%以上,达到25帧/s的高精度实时图像处理要求.

关 键 词:高精度图像处理  硬件加速  嵌入式  DSP  PCI  可编程逻辑器件  硬件加速  精度  实时图像处理系统  系统设计  Hardware  Acceleration  Based  System  速度提高  算法  验证  系统处理  并行处理  流水线  设计实现  软件  封装  数据传输  控制通路  数据通路  主机
文章编号:1671-4598(2007)03-0387-03
收稿时间:2006-05-08
修稿时间:2006-06-01

High-precision Real-time Image Processing System Based on Hardware Acceleration
Wang Zhe,Zhang Xiaolin,Liu Rongke,Wang Dong,Zhao Guangsheng.High-precision Real-time Image Processing System Based on Hardware Acceleration[J].Computer Measurement & Control,2007,15(3):387-389.
Authors:Wang Zhe  Zhang Xiaolin  Liu Rongke  Wang Dong  Zhao Guangsheng
Affiliation:School of Electronic and Information Engineering, BUAA, Beijing 100083, China
Abstract:Higb-precision real time image processing is widely used in many fields, and such an image processing system based on hardware acceleration is presented. In the system, a float -point embedded assistant processing unit is used to accelerate the high precision processing, while data and control interfaces between assistant processing unit and the host give facilities for the image data being transferred rapidly and software package of the unit, and pipe-line parallel processing is achieved via the proposed software design. It is experimentally shown that, as for the algorithm we employed the speed of image processing is evidently improved for more than 300% compared to host pro cessing, and high- precision real -time processing of 25 frame/sec is achieved.
Keywords:high- precision image processing  hardware acceleration embedded  DSP  PCI  PLD
本文献已被 CNKI 维普 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