Abstract: Pointing at the pixel-level image fusion for multi-source digital image, this paper gives a new method for image fusion based on wavelet transform. The choice of different fusion rules is decided by different characteristics of the frequency domain. The low-frequency coefficients are given based on the information of the image edge. The high-frequency coefficients are given based on the combination of window and the fractal dimension. Experimental results demonstrate that the image fused by the new method is better than using traditional approaches both in visual effect and objective evaluation, which provides a new method for the research of image fusion.

Key words: wavelet transform, image fusion, fusion rules

摘要： 针对多源像素级数字图像融合，给出了一种基于小波分析的融合新方法。根据不同的频率域特点选择不同的融合规则，对低频系数给出了基于边缘信息的融合规则，对高频系数给出了基于窗口与分形思想相结合的融合规则。实验结果表明，新方法得到的融合图像无论在视觉效果还是在客观评价方面，均优于传统的融合方法，这为图像融合算法的研究提供了一种新方法。

关键词: 小波分析, 图像融合, 融合规则