基于二维统计经验模态分解的三维面形测量

Three-dimensional shape measurement based on bidimensional statistical empirical mode decomposition

提出一种基于二维统计经验模态分解(BSEMD,bidimensional statistical empirical mode decomposition)的三维面形检测方法。BSEMD在构造上下包络时不是直接采用插值的方法而是引入了平滑过程,因此比传统的经验模态分解(EMD)方法有更好的抑噪效果。首先对采集到的结构光变形条纹进行BSEMD后,去掉背景分量可得到含有被测物体高度信息的基频分量;然后结合希尔伯特变换得到调制相位;最后通过相位和高度的映射公式计算出物体的高度。计算机模拟和实验都验证了本文方法的有效性。

A new three-dimensinal shape measurement method is proposed based on hidimensional statis- tical empirical mode decomposition （BSEMI））. When constructing two-dimensional upper and lower en- velopes,BSEMD uses a smoothing process instead of an interpolation. The BSEMD method has better noise inhibition performance than conventional BEMD method because of the smoothing sifting proce- dure. After the deformation fringe of structured light is decomposed by BSEMD, the useful fundamental components including measured object＇s height information can be obtained by removing the background components. And then we can obtain the modulated phase distribution by Hilbert transform. The height of measured object can be retrieved by mapping the phase and height. The simulations and experiments prove the validity of this method.