首页 | 本学科首页   官方微博 | 高级检索  
     

金字塔双层动态规划立体匹配算法
引用本文:赵杰,于舒春,蔡鹤皋.金字塔双层动态规划立体匹配算法[J].控制与决策,2007,22(1):69-72.
作者姓名:赵杰  于舒春  蔡鹤皋
作者单位:哈尔滨工业大学,机器人研究所,哈尔滨,150080
基金项目:国家高技术研究发展计划项目(2001AA422250).
摘    要:针对控制点修正的动态规划立体匹配算法存在控制点求取时阀长、实时性差的问题.提出一种金字塔双层动态规划立体匹配算法.采用金字塔算法求取低、商分辨率图像,然后分别在低、商分辨宰图像上求取候选控制点集和最终控制点集,并用最终控制点集修正商分辨率图像上的动态规划立体匹配.由干候选控制点集的求取在低分辨率图像上进行,算法用时大为减少.实验证明,此算法匹配率商、速度快.

关 键 词:控制点  动态规划  立体匹配  视觉导航
文章编号:1001-0920(2007)01-0069-04
收稿时间:2005-10-11
修稿时间:2005-10-112005-12-05

Pyramid two-level dynamic programming stereo matching algorithm
ZHAO Jie,YU Shu-chun,CAI He-gao.Pyramid two-level dynamic programming stereo matching algorithm[J].Control and Decision,2007,22(1):69-72.
Authors:ZHAO Jie  YU Shu-chun  CAI He-gao
Affiliation:Robot Research Institute, Harbin Institute of Technology, Harbin 150080, China
Abstract:To the problems of dynamic programming stereo matching algorithm modified by ground control points(GCPs),a pyramid two-level dynamic programming stereo matching algorithm is proposed.The low-resolution images and high-resolution images are gained by pyramid algorithm.Then,candidate GCPs and ultima GCPs are obtained from low-resolution and high-resolution images.Dynamic programming stereo matching algorithm in high-resolution images is directed by ultima GCPs.Total time of this algorithm is reduced because candidate GCPs are gained in low-resolution images.Experiments show that the algorithm has a high maching-rate and a fast speed.
Keywords:Ground control points(GCPs)  Dynamic programming  Stereo matching  Vision-navigation
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《控制与决策》浏览原始摘要信息
点击此处可从《控制与决策》下载全文
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