基于知识的机器人视觉测量分类与进展-控制与决策.PDFVIP

第25 卷 第4 期 控 制 与 决 策 2010 年 4 月 Vol. 25 No. 4 Control and Decision Apr. 2010 文章编号: 1001-0920 (2010) 04-0481-06 基于知识的机器人视觉测量分类与进展 徐 德, 谭 民 (中国科学院自动化研究所复杂系统与智能科学重点实验室，北京100190) 摘 要: 提出了基于知识的视觉测量的观点, 将传统的立体视觉看作一种基于摄像机知识的视觉测量, 并对当前其 他的视觉测量方式按所采用知识的不同进行了分类. 分别简要介绍了基于点、线、面、统计特性等目标知识的视觉测 量, 讨论了基于运动知识的视觉测量和基于环境知识的视觉测量. 在对视觉测量所采用知识的表达和获取进行初步 探讨的基础上, 给出了视觉测量的几个值得深入研究的发展方向. 关键词: 视觉测量；仿人视觉；机器人；知识获取；知识表达 中图分类号: TP242 ；TP212 文献标识码: A Development and classification of knowledge-based visual measurement for robots XU De , TAN Min (Key Laboratory of Complex Systems and Intelligent Science, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China. Correspondent: XU De, E-mail: xude@compsys.ia.ac.cn) Abstract: A new view point of knowledge-based visual measurement is proposed, which takes the traditional stereovision as a kind of visual measurement based on cameras’ knowledge. The currently works on visual measurement are classed to different categories based on the different kinds of knowledge used. The visual measurement methods based on the object’s knowledge such as point, line, plane, and statistic characteristics are shortly introduced. Then the methods based on motion knowledge and environmental knowledge are discussed. The knowledge expression and acquirement for visual measurement are concisely commented. Finally, several potentially developmental topics that are valuable to be further investigated are suggested. Key words: Visual measurement ；Humanoid vision ；Robot ；Knowledge acquirement ；Knowledge expression [1, 2] 1 传传传统统统的的的视视视觉觉觉测测测量量量