精确标定关节臂视觉检测系统手眼关系  

Accurate calibration of AACMM visual detection system hand-eye relationship

作　　者：王金桥[1] 段发阶[1] 汪润[1] 

机构地区：[1]天津大学精密测试技术及仪器国家重点实验室,天津300072

出　　处：《计算机工程与应用》2015年第21期225-229,共5页Computer Engineering and Applications

基　　金：国家高技术研究发展计划(863)(No.2013AA102402);国家自然科学基金(No.51275349)

摘　　要：在关节臂视觉检测系统中,视觉传感器与关节臂末端的手眼关系标定的准确性直接影响到系统的测量精度。推导了关节臂手眼问题的数学关系模型,提出了一种基于遗传优化算法的双阶段手眼标定的新方法,即先采用常规方法求取手眼变换矩阵主要参数的初始值,在初始值的基础上利用遗传算法优化手眼标定模型的所有参数,建立遗传优化算法的手眼关系数学模型。通过实验论证了该方法的优良性和可行性,对一般手眼问题具有一定的可移植性,系统测量精度优于40μm。In the AACMM（Articulated Arm Coordinate Measuring Machine）visual inspection system, the accuracy of measurement system has been directly affected by the relationship of hand-eye calibration between vision sensor and the end of AACMM. The mathematical relationship of AACMM hand-eye problem has been derived in this paper, and a new method has been proposed, which is a two-stage hand-eye calibration algorithm based on genetic optimization. The initial value of main parameters of hand-eye transformation matrix are got by using conventional methods, and all parameters of hand-eye calibration model on the basis of the initial value are optimized by using genetic algorithm. A mathematical model of the hand-eye relationship genetic optimization algorithm is established. Experiments demonstrate excellence and feasi-bility of this method and the general problem of hand-eye has some portability, and the accuracy is better than 40μm.

关 键 词：关节臂 线激光 视觉传感器 手眼标定 遗传算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]