计算机工程 ›› 2011, Vol. 37 ›› Issue (14): 143-145.doi: 10.3969/j.issn.1000-3428.2011.14.047

• 人工智能及识别技术 • 上一篇    下一篇

双目主动视觉监测平台下的目标识别

孔令富 1,连秀梅 1,赵立强 2   

  1. (1. 燕山大学信息科学与工程学院,河北 秦皇岛 066004;2. 河北科技师范学院,河北 秦皇岛 066004)
  • 收稿日期:2010-12-15 出版日期:2011-07-20 发布日期:2011-07-20
  • 作者简介:孔令富(1957-),男,教授、博士、博士生导师,主研方向:智能控制,机器人技术,并行分布式系统;连秀梅,硕士研究生;赵立强,教授、博士
  • 基金项目:
    国家自然科学基金资助项目(60975062);河北省教育厅自然科学研究计划基金资助项目(Z2009115)

Object Identification in Binocular Active Visual Monitoring Platform

KONG Ling-fu 1, LIAN Xiu-mei 1, ZHAO Li-qiang 2   

  1. (1. School of Information Science and Engineering, Yanshan University, Qinhuangdao 066004, China;2. Hebei Normal University of Science & Technology, Qinhuangdao 066004, China)
  • Received:2010-12-15 Online:2011-07-20 Published:2011-07-20

摘要: 利用对目标旋转、尺度变化、视角变化等具有稳定性的尺度不变特征变换(SIFT)算法,提出一种适用于基于圆轨的基线可调双目主动视觉监测平台的目标识别方法,通过离线建立物体的多侧面SIFT特征点数据库,将三维空间的目标转换为二维特征描述,利用二维特征描述实现三维空间目标的识别,以提高匹配识别效率。实验结果表明,该方法能实时准确地识别目标。

关键词: 尺度不变特征变换, 目标识别, 二维特征描述, 双目主动视觉, 特征点数据库

Abstract: Scale Invariant Features Transform(SIFT) algorithm is stable to rotating, scale changes and visual angle changes of the target, so this paper proposes an object identification method for a new visual monitoring platform——the baseline based on circle track adjustable binocular active vision platform. By establishing the off-line database of multifaceted SIFT features, it converts three-dimensional object into two-dimensional characterization, and uses it to realize object recognition three-dimensional space, so that the matching and recognition efficiency is improved. Experimental results show that the method can identify objects real time and accurately.

Key words: Scale Invariant Feature Transform(SIFT), object recognition, two-dimensional characterization, binocular active vision, feature point database

中图分类号: