多传感器集成测量系统的数据对齐方法研究

反求工程中，多传感器集成测量系统对于提高实物数字化的效率和精度有着重要意义。本文对多传感器集成测量系统中不同数字化方式所获数据的对齐和融合进行了研究，并提出了一种不依靠简单几何体作为基准的自动对齐算法。该算法首先通过计算点集曲率来辅助选取3对近似对应点，求解坐标变换矩阵实现初始对齐；然后对非接触扫描点云进行三角剖分，以接触式测量所获点集向对应三角片的投影寻找对应点，然后进行坐标变换迭代求解，从而实现了不存在对应点的点集之间的对齐；最后包围盒求交法实现了数据的融合。实验结果表明，该算法行之有效，运行效果良好。

Data Registration of Multiple-sensor Integrated Measuring System

Multiple-sensor integrated measuring system (MIMS) is of great importance to improve the efficiency and the precision of digitization in reverse engineering. Registration and data merging of different digitization result from MIMS are studied in this paper, and an automated registration method is proposed too, in which primitive artifacts are not required as datum. In the algorithm, curvature is calculated firstly according to the point and its neighbor points, and then three pairs of points which have similar curvature are selected to compute transform matrix for initial registration. Afterwards, in order to get corresponding points, point set of contact measurement are projected to corresponding triangular mesh constructed from results of non-contact measurement, and the transformation can be achieved by iterative method. Finally intersection of bounding box is utilized to merge data into one set. Several experimental results are presented to demonstrate the feasibility and performance of the proposed algorithm.