船体分段三维数字化测量数据配准

The data registration of 3-D digital measurement for hull blocks

船体分段测量数据与CAD模型的配准,直接影响船体分段的制造几何误差分析和制造精度评价.针对船体分段装配工艺以及测量数据点集的分布特点,提出了一种基于单位四元数的Levenberg-Marquardt非线性迭代的最优配准算法.首先用一定的基准(肋骨对齐)对测量的角点数据进行初始变换,然后用单位四元数算法对初始变换后的数据和CAD模型进行粗配准,使变换后的船体分段测量数据与CAD模型数据处于小误差状态,最后采用基于L-M非线性迭代法求全局最优配准参数.实验结果表明,该算法能很好地反映船体分段建造中的实际情况,为船体分段的后续装配提供指导性的依据.

Manufacturing errors caused by lack of manufacturing precision for hull blocks are directly related to registration errors in hull block measurement data and CAD models. After examining the assembly process for hull blocks and the distribution features of measurement data, a Levenberg-Marquardt （LM） optimal registration algorithm based on unit quaternion was proposed. The algorithm first made an initial transformation of measured corner data by reference （rib alignment）. It then made coarse registration with the unit quaternion algorithm in order to minimize errors between the transformed block measurement data and the CAD model. Finally the global optimal registration parameters were derived by the non-linear iterative method based on LM. Experimental results showed that the algorithm reflects the actual situation of hull block construction very well and provides an instructive basis for subsequent assembly.