论文部分内容阅读
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 three-dimensional surface detection method based on two-dimensional statistical empirical mode decomposition (BSEMD) is proposed. BSEMD constructs the upper and lower envelopes not directly using the interpolation method but introduces the smoothing process, so it has better noise suppression effect than the traditional empirical mode decomposition (EMD) method. First of all, after the BSEMD is extracted, the background components are removed to obtain the fundamental frequency components containing the height information of the measured object. Then the phase modulation is obtained by combining the Hilbert transform. Finally, the phase and height mapping formulas are used to calculate The height of the object. Computer simulation and experiments verify the effectiveness of the proposed method.