摘要
讨论了基于双焦的单目立体成像模型,分析了在共轴模型下CCD成像离散性而导致的深度计算误差与镜头焦距及物点空间位置的关系,当镜头焦距增加时,可精确恢复深度增大,当物点距镜头光轴距离较远,即物点位于视场边缘时,深度计算误差较小;并根据双焦成像特点提出了相应的特征匹配方法,即采用与两焦距倍数相关的特征提取算子,并对匹配结果进行相应的深度计算及插值;文章最后给出了理想双焦图像对的相应实验结果,并探讨了实验误差的形成原因。
This paper discusses the model of monocular stereo system based on bifocal imaging, analyses the relationship between the depth calcu lation error in the coaxial model caused by the discrete CCD imaging and the focus or the position of the object point. It shows that when the camera fo cus increases, the accurate depth acquired increases, when the object point is further away from the optical axis, the depth error becomes less. Accord ing to the characteristic of the bifocal imaging, relevant feature matching algorithm are presented, in which the parameter of the feature detection is de cided by the relationship between the two focus. According to the match result we can calculate the depth map. The experiment results of the ideal bi focal image pairs is shown in the end of the paper, and the reason of the experiment error is analyzed.
出处
《计算机测量与控制》
CSCD
2008年第9期1316-1318,1321,共4页
Computer Measurement &Control
基金
国家自然科学基金资助项目(60472026)
关键词
单目立体视觉
深度误差
特征匹配
深度插值
monocular stereo vision
depth error
feature matching
depth interpolation