基于Meta-face Learning的工件定位算法
Workpiece Localization Algorithm Based on Meta-face Learning

提出了一种包含自由曲面特征的工件定位的Meta-face Learning(MFL)算法。利用基于字典学习的图像稀疏表示方法,在交替迭代优化的基础上,通过逐次修正Euclidean变换矩阵的列向量更新测量点到名义工件模型的位姿变换,确定工件坐标系相对于测量坐标系的位姿。设计了两个自由曲面验证了本文算法,并通过与现有算法的比较说明了其具有较高的计算效率和定位精度。
A Meta-face Learning(MFL) localization algorithm for workpiece including free form surfaces is presented in this paper. With the method of image sparse representation based on dictionary learning with alternating iteration optimization, the transformation is updated from the measurement data to ideal geometric model with aligning the column vector of euclidean transformation matrix individually, and then the configuration of design frame is determined with respect to measurement frame. Two free form surfaces are used as examples to validate the effectiveness of the proposed algorithm and the comparative analysis with existing algorithms demonstrates the high computational efficiency and location accuracy of the algorithm