基于最小二乘原理的点云数据配准

Point Cloud Data Registration Based on Least Squares

摘要: 基于最小二乘原理，本文研究了点云三维坐标转换的模型及其迭代解法，介绍了一种不求旋转角度而将旋转矩阵的元素作为未知参数进行求解的方法，将原本的七参数模型转换成十三参数模型，介绍了求解参数的模型和解法。弥补了传统七参数法在面对大旋转角时的不足，该方法能够解决任意旋转角的三维坐标转换。

Abstract: Based on the principle of least squares, this paper studies the point cloud three-dimensional coor-dinate transformation model and iterative solution method. It introduces a method that solves the rotation matrix’s elements as an unknown parameter without seeking the rotation angle, and converts the original seven-parameter model into a thirteen-parameter model, and the model and solution to solve the parameters are introduced. It compensates for the shortcomings of the tradi-tional seven-parameter method in the face of large rotation angle. This method can solve the three-dimensional coordinate transformation of arbitrary rotation angle.