北京航空航天大学学报 ›› 2009, Vol. 35 ›› Issue (11): 1224-1227.

• 论文 • 上一篇    下一篇

改进的SSDA并行算法及其在TEM中的应用

胡 凯, 张新宇, 杨 锐, 刘德卿   

  1. 北京航空航天大学 计算机学院, 北京 100191
  • 收稿日期:2008-10-09 发布日期:2010-09-16
  • 作者简介:胡 凯(1963-),男,湖南长沙人,副教授,hukai@buaa.edu.cn.
  • 基金资助:

    航空科学基金资助项目(2007ZC51032); 国家863计划资助项目(2007AA01A127)

Parallel algorithm based on improved SSDA in TEM

Hu Kai, Zhang Xinyu, Yang Rui, Liu Deqing   

  1. School of Computer Science and Technology, Beijing University of Aeronautics and Astronautics, Beijing 100191, China
  • Received:2008-10-09 Published:2010-09-16

摘要: 介绍了一种把图像匹配中的序贯相似性检测算法(SSDA,Sequential Similarity Detection Algorithm)应用到地形匹配中的方法,把地形匹配系统中的数字高程模型、高程数据和实时剖面数据分别视为SSDA算法中的搜索图、灰度值和模板,同时提出了动态门限序列和分组选取随机点的方法,最后将改进的SSDA算法进行并行化并用消息传递接口(MPI, Message Passing Interface)实现.实验结果表明,改进的SSDA算法能有效地提高地形匹配的速度和计算精度,而相应的并行程序也能取得较好的加速比,从而一定程度上解决了传统的地形匹配算法时间复杂度高、实时性差的问题.

Abstract: A new terrain elevation matching (TEM) method using sequential similarity detection algorithm (SSDA), which is an image matching algorithm was presented, while regarding the digital elevation model, height value and real-time profile data in TEM as searching image, hue value and template in SSDA respectively. The dynamic threshold sequence and the method of selecting random points in groups were described. A parallelism algorithm based on improved SSDA was designed and implemented with message passing interface(MPI). The experiment results show that the improved SSDA algorithm can effectively increase the matching speed and the precision, and the corresponding parallel program can also get good speed-up ratio. Therefore the problem of high time complexity and lack of real time feature in traditional TEM was solved to some extent.

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发