A new multi-focus image fusion algorithm based on BEMD and improved local energy

A multi-focus image fusion method, which is based on bidimensional empirical mode decomposition (BEMD) and improved local energy algorithm, is presented in this paper. First, the source image is decomposed by BEMD. Then maximum criterion combined with weighted average fusion rule based on local energy is applied to bidimensional intrinsic mode function components of the corresponding frequency segment. If the phase of the bidimensional intrinsic mode function coefficients decomposed by BEMD on two source images is the same, the local energy maximum criterion is used in the frequency coefficients of the fused image; else, if the corresponding phase is opposite, bidimensional intrinsic mode function coefficients of the fused image is determined by a weighted average method based on local energy. Finally, the fusion result is obtained by the inverse BEMD transformation on the fusion coefficient. Simulation shows that the proposed algorithm significantly outperforms traditional methods such as the maximum criterion, weighted average, and wavelet fusion rules.