关键词: 集合经验模态分解, 游程检测法, 奇异谱分析, 曲线光顺算法

Abstract: Aiming at the curve smoothing,a curve smoothing algorithm based on empirical mode decomposition,run test refactoring and singular spectrum analysis noise reduction was proposed.Three variables x,y and z on the spatial discrete digital curve were regarded as three one-dimensional digital signals.Ensemble Empirical Mode Decomposition (EEMD) decomposition was carried out for each variable's digital signal sequence.Then,the runs test method was used to reconstruct all components of each variable into low-frequency components and high-frequency components.Singular Spectrum Analysis (SSA) was used to reduce the noise of the reconstructed high frequency components.The high frequency and low frequency components after noise reduction were reconstructed to obtain the curve after smoothing.Experiments showed that the smoothing effect of this algorithm was better than that of Empirical Mode Decomposition (EMD) method and curvature method.The average curvature of the proposed algorithm,EMD method and curvature method were 0.0893,0.0919 and 0.1112 respectively.

Key words: ensemble empirical mode decomposition, runs test method, singular spectrum analysis, curve fairness algorithm