Abstract
© 2014 IEEE.Empirical mode decomposition (EMD) is a powerful tool for analysis of non-stationaryand nonlinear signals, and has drawn significant attention in various engineeringapplication areas. This paper presents a new bidimensional EMD based on the adaptive anisotropic triangulations. Specifically, we define the local mean surface of the data, which is a key step in bidimensional EMD, by a locally weighted mean filter with variable window sizes that are determined by the adaptive anisotropic triangulations. Numerical experimentsshow that the proposed method achieves effective empirical mode decomposition for 2D signals.