基于投影和主成分分析的偏心检测

Eccentricity measurement based on projection and principal component analysis

在生产实践中,管类工件的内部结构之中常常存在微小的偏心误差。偏心误差对工件的工作性能影响很大,因此,能否准确、快速地检测出偏心误差在生产实践中意义重大。本文介绍了一种基于投影数据和主成分分析的偏心检测方法。截取工件中任意两个截面,利用Radon变化求得它们在一定投影角度下的投影。分别计算投影的第一主成分,并以第一主成分之间的欧氏距离表征偏心误差。如果欧氏距离不等于零,则证明工件存在偏心误差。通过仿真实验证明该方法能检测出0.1%的微小偏心,并且抗噪性能较好。

In the productive practice,the internal structures of pipe parts and pipe products often have minimal eccentric error. Eccentric error gets great effect to the performance of workpiece. It is important to measure the eccentric error exactly and fast in the productive practice. This paper introduces a kind of methods to measure eccentric error,which based on projection data and principal component analysis. Intercept any two sections of workpiece,and collect their projection date from Radon transformation. Compute the first principal components of projection respectively,then use Euclidean distance to characterize the eccentric error. If the Euclidean distance is not equal to zero,the workpiece exist eccentric error. The simulation experiment proves that this method can test minimal eccentric error of 0.1%,and it has good anti-noise performance.