Abstract: In order to improve the efficiency of image fusion, aiming at the limitation of current image fusion methods, the author proposed a multifocus image fusion method based on wavelet transform. Firstly, the optimal wavelet decomposition level was determined, and wavelet transform was used to decompose image. Secondly, considering the characteristics of low frequency subband and highfrequency subband, the local average gradient criterion was chosen as the fusion rule of highfrequency subband, and the average value of three coefficients was used as the fusion rule of lowfrequency subband. Finally, the validity of image fusion was tested by simulation experiment. Experimental results show that the proposed method achieves more ideal image fusion results, and improves the quality of fused image. The fusion effect is better than the contrast image fusion method.

Key words: image fusion, fusion rule, multi-focus image, wavelet transform, peak signal to noise ratio