摘要
为了直观反映双目视觉系统各参数标定误差对系统定位精度的影响,基于针孔摄像机成像模型和概率统计理论,运用蒙特卡罗法对轴线平行结构的系统定位误差进行仿真分析。针对视觉定位模型求解空间点表达式复杂的问题,为便于仿真,根据成像模型参数的相互独立性,提出了一种单参数分析定位误差的方法。利用该方法进行的仿真结果及其实验验证表明:内参数中,主点(u0,v0)的标定误差对系统X、Y方向定位精度影响大,焦距fu的标定误差对系统Z方向定位精度影响大;外参数中,旋转角β和基线t1对X、Y、Z3个方向定位精度的影响都较大。该结论对双目视觉系统设计及参数标定具有一定的参考价值。
The calibrated parameters that caused errors in parallel-axes binocular stereoscopic measurement system were studied. Based on the pinhole camera model and the theory of probability and statistics, the Monte-Carlo method was used in simulating and analyzing the errors. In the view of complex algebraic expression of 3D point from the positioning model of binocular vision, according to the independence of imaging-model-parameters, a method of single-parameter analysis positioning errors was proposed to facilitate the simulation. The simulation result and experiment show that among the intrinsic parameters, the calibration error of the principle point （ uο ,vο ） has a significant influence on the positioning accuracy on the X and Y direction, and the calibration error of focus fu has an obvious influence on the positioning accuracy on the Z direction. On the other hand, among the extrinsic parameters, both the rotation angle β and the baseline t1 have a visible influence on the positioning accuracies on all the three direction. The conclusion can be taken as a reference for the design and calibration of binocular vision system.
出处
《电光与控制》
北大核心
2009年第12期65-69,共5页
Electronics Optics & Control
基金
国家十一五科技支撑计划(2006BA103A16)
北京市科技计划(H060720050230)
关键词
双目视觉
针孔相机模型
蒙特卡罗法
标定误差分析
binocular vision
pinhole camera model
Monte-Carlo method
calibration error analysis
