摘要
准确地获取实体的三维信息一直是精密测试、CAD/CAM、机器人导向等领域中的重要课题。本文提出了一种新的基于粘性目标约束和外极约束的自由曲面立体视觉测量方法 ,这种方法有效解决了立体视觉中的图像匹配问题。外极约束将可能的侯选点限制于直线分布 ,引入粘性目标用以确定光栅条纹的级次。这种方法极大地减小了图像匹配的运算量 ,同时减小了错误匹配的概率。实验表明 ,采用光栅投影主动视觉传感器能够高效准确地测量自由曲面的三维轮廓。
It is a very important task to obtain accurate 3 D information in precision inspection,CAD/CAM,robot navigation and other such fields.This paper presents a new method of stereovision profilometry based on adherent mark constraint and epiploar constraint(AMC & EC).With this method,the correspondence problems are easily and precisely settled.Epipolar constraint is applied to limit candidate matching point to a line distribution,and the adherent mark constraint is used to tell the grating index.This method saves computational time greatly and miscorrespondences are reduced at the same time.Using grating projection and stereo matching based on these constraints,free form 3 D profiles can be measured efficiently and accurately.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2002年第3期324-327,330,共5页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金资助项目
(No 50 0 750 63)
教育部科学技术重点项目 (No991 4 0 )
关键词
三维轮廓术
立体视觉
粘性目标约束
外极约束
图像匹配
D Profilometry Stereovision Adherent mark Constraint Epipolar constraint Image registration