基于稳健滤波技术的测量异常值处理的研究

The Study of Processing to Outlier in Measured Data Based on Robust Filtering

本文基于稳键滤波的 M估计技术 ,提出了一种全新的、对大型旋转机械的表面参数测量中出现的异常值的处理方法 ,并给出了数学模型和实验结果。经过此方法处理后的测量数据 ,可以保证在时间顺序上不会发生断裂 ,而且消除了异常值的影响 ,取得了令人满意的结果并为数据的进一步处理、存储和显示做好准备工作。

Based on M Estimate technology of robust filtering, this paper proposes an entirely new method for the processing to the outlier, which exist in the surface parameter measurement of large rotary machine. It also presents the mathematic model and result of experiment. After the processing using this method, the data avoid to be parted in timely sequence, and the effect of the outliers in the data is eliminated. This process outcome is satisfied and well prepares for the further processing, storing and displaying of the data.