基于点拟合的大尺寸视觉测量精度评价方法  

Evaluating the accuracy of large scale vision measurement by fitting point

作　　者：杨剑[1,2] 吕乃光[1,2] 董明利[2] 燕必希[2] 

机构地区：[1]北京邮电大学电子工程学院,北京100876 [2]北京信息科技大学光电信息与通信工程学院,北京100192

出　　处：《华中科技大学学报（自然科学版）》2010年第9期22-25,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(50675015);北京市自然科学基金资助项目(4102020);现代测控技术教育部重点实验室资助项目;北京市重点实验室开放课题基金资助项目(KE20091123208)

摘　　要：为了现提高场评价视觉测量系统的测量精度,提出以距离约束为基础进行点的拟合,利用拟合点估计偏差和不确定度进行评定的方法,并将拟合点不确定度定义为点的离散度.在视场内建立多个基准距离,利用距离和坐标系的选取无关的性质,在视觉测量系统中基于距离约束进行点的拟合,利用拟合点的不确定度来表示整个视觉测量的不确定度.实验中,在置信水平为95%的条件下,测量的87个点中只有1个超出了极限偏差,拟合点离散度估计和统计结果符合较好.A distance constraint-based fitting point is proposed to evaluate the accuracy of a large scale vision measurement system and reconcile the measurement result through national standards. The measurement accuracy evaluation is described by the variance of the fitting point and named ＂fitting point variance＂. Firstly, many line segments were considered as true value. Use the fitting uncertainty to express the whole the measuring uncertainty. In the experiment, the distance between the control target points are measured by the laser tracker and are considered as true value. Firstly, the measuring point errors are estimated. In confidence level of 95 %, 87 points were measured and only one point deviation exceeded the estimation. The measurement uncertainty estimation of fitting points indicates that the result is consistent with the statistical results.

关 键 词：视觉测量 现场大尺寸测量 精度评价 距离约束 点拟合 

分 类 号：O435[机械工程—光学工程] TP394.1[理学—光学]