多变量总体最小二乘在点云拼接中的应用

Application of Multivariable Total Least Squares in the Registration of Point Clouds

在三维激光扫描仪使用过程中,为了减小点云拼接时的误差问题,本文利用同方差多元变量的EIV(Errors In Variables)模型及总体最小二乘的方法解决三维空间点的相似变换,较传统的迭代算法计算空间坐标转换的方法,具有非迭代性、可靠性和计算过程中的简便性。最后,利用实际工程案例对非迭代算法的有效性进行了验证。

In the use of process of 3D laser scanner, this paper uses new TLS -type algorithm in a multivariate homoscedastic EIV（Errors in Variables）model and the total least 0 -square（TLS） methods to solve similarity transformation parameters of three -dimensionalspace points in order to reduce the error of point cloud registration .Comparing with the traditional iterative algorithm in spa -tial coordinate transformation calculation , it is without iteration and has the advantages of simplicity , reliability in the computingprocess.Finally, the validity of non iterative algorithm is verified in the actual project cases .