Abstract:

For complex structure blanks with small machining allowance, small deviations may result in a shortage of allowance during the machining operation. In order to achieve optimal allocation of machining allowance, a hierarchical and fast algorithm of constrained registration between the dense scanned data of blank parts and the nominal finished part was presented. Firstly, with the aid of an elaborately built segmentation guide, each surface to be machined was separated from the point cloud by using points' local differential information. The minimal allowance constraint and an initial data set, which will be involved in the following registration process, were then assigned to each segmented surface respectively. The initial registration datasets were aligned to the design model by using multiplier methods, and then the measured data points that violate the allowance constraint were added iteratively into the current registration dataset until all the scanned points satisfied the allowance constraints. Experiments show that the proposed algorithm can achieve allowance-constrained registration with great efficiency.

Key words: allowance constraint;registration, point cloud segmentation, measurement data

摘要：

针对复杂结构毛坯的小余量数控加工中很容易出现局部加工余量不足的问题,基于毛坯表面的点云测量数据,提出了一种层次化快速约束配准算法,从而实现加工余量的优化分配。首先通过建立一个知识向导模型并应用局部微分信息,自动分离出毛坯上需要分配余量的各待加工表面,并分别施加最小余量约束,在确定各待加工表面上参与约束配准优化的初始点集的基础上,采用广义乘子法将初始配准点集与设计数模进行约束匹配;然后采用迭代方法将不满足余量约束的测点追加到当前配准点集中,直到所有待加工表面的测点满足余量约束。实验结果表明,该配准算法在最大程度满足余量约束的前提下,显著提高了点云数据的配准效率。

关键词: 余量约束, 配准, 点云分块, 测量数据