小型回转点云的配准研究

Registration Study of Small Turning Point Clouds

针对回转半径较大的小型回转点云配准时其回转特征不明显问题,提出一种先利用最小二乘法计算出点云的拟合平面并将所有点投影到该平面上,再利用滚球算法计算该二维点云数据的边界,并对边界点云进行3次B样条拟合,最后根据B样条曲线的变化特点,找出特征点,并利用特征点进行配准。对于回转半径较大且尺寸较小的回转点云,由于其回转特性不是非常明显,因此完全依靠其回转特性进行配准,会有一定的难度,所以需要利用其他一些特征进行配准。最后通过三坐标扫描仪获取物体的外表面形状数据,然后进行配合验证,结果表明可以准确获取特征点。该方法可以有效地对小型点云进行配准。

Aiming at the problem that small turning point clouds with larger turning radius do not have obvious rotary characteristics, this paper proposed to calculate the point cloud of fitting plane using the method of the least squares first and to project all the points to the plane. Then by the method of ball algorithm, the border of the clouds could be calculated, and though three times B-spline curve fitting, the feature points of the curve could be found and further to be used in registration. This article got the primary points by the scanner coordinate, then verified with the experimental results. The experimental results show that this method can be effective for small point clouds registration.