数字化测量中的最佳配准问题

Digital Measurement Best-Fitting

针对数字化测量中的最佳配准这一问题,提出了一种快速准确的配准方法.该方法首先建立配准数学模型,然后借鉴LM(Levenberg-Marquardt)算法与奇异值分解的优点,实现了测量数据与CAD模型之间的配准,最后为验证配准矩阵的有效性,提出了一种针对匹配结果的确定性分析方法.实验验证该配准方法能较好地实现测量数据与CAD模型间的配准,进而为后续的模型检测与模型评估奠定基础.

A registration method is presented for best-fitting in digital measurement. First the mathematical model is established. Then the virtues of LM （Levenberg-Marquardt） algorithm and singular value decomposition are integrated, and then measurement data and CAD model are matched. Finally a confirmation analysis method is proposed for validating the validity of registration matrix. Experimental results show the method can realize the registration between measurement data and CAD model and it establishes the base for the following model testing and model evaluation.