白车身小样本抽样下的尺寸数据处理

Process of the Small Nmber Sampling Data of Boby in white

本文针对白车身小样本采样下的CMM (三座标 )数据 ,提出新的突变点剔除办法 -平均值法 ,利用数字信号处理中用到的建立在多项式拟合基础上的改造后的简单整系数滤波器将白车身的尺寸数据的趋势项与波动项的分离 ,以得到较为真实的 6Sigma值 。

A new method is presented to process the small number sampling data of CMM(Coordinator Measuring Machine) in Body in White. First, a sample value is decided to whether be a mutation or not by comparing its value to the average of its three neighborhood at each side. Second, to separate the original data into trend and variation sequences by a modified integer coefficient filter based on the polynomial fit in the area of digital signal processing. Through this way,true 6Sigma value is gotten to evaluate the BIW dimension quality.