Study of Image Fusion Based on Multiwavelet Transform and Fractal Dimension

PU Jie-xin,TANG Guo-liang,HUANG Xin-han,ZHANG Zhi-yong
DOI: https://doi.org/10.3321/j.issn:1002-8331.2006.32.025
2006-01-01
Abstract:Multi-wavelet is an extension from wavelet theory,and has several particular advantages in comparison with scalar wavelets on image processing.Multiwavelet analysis can offer a more precise way for image analysis than wavelet multi-resolution analysis.Because images in different levels of image multiwavelet decomposition were self-similar,which was the foundation of fractal,a new image fusion algorithm was presented based on image multiwavelet decomposition and fractal dimension.Different source images were separately decomposed by multi-wavelet transform with different scales.Corresponding-level high frequency images were merged,with fractal dimension as weight.Corresponding-level lowest frequency images were merged,with energy as weight.An experiment was given using multi-focus images,visual light and infrared image separately as two source images.Experiment results show the algorithm is feasible.
What problem does this paper attempt to address?