基于小波变换的自适应图像融合算法

在分析已有小波图像融合方法的基础上,针对高、低频融合规则的选择问题,提出了一种基于小波多分辨率分解的图像融合算法。该算法对小波分解后的低频子图像采用基于主成分分析的低频融合规则进行融合,而对高频子图像采用系数绝对值取最大和基于局部均值方差最大化的融合规则进行融合。实验结果表明,该方法提高了融合图像包含的信息量,最大可能地消除了局部对比度极性反转的情况,明显地增强了融合图像的清晰度,而且很好地保留了源图像中的边缘细节。

Self-Adaptive Image Fusion Algorithm Based on Wavelet Transform

Based on the analysis of existing wavelet image fusion method and with an aim at the selection of high and low frequency fusion rules,the paper presents a new image fusion algorithm based on wavelet multi-resolution decomposition.A low frequency fusion rule based on the principal component analysis(PCA) is used to fuse the low frequency sub-images decomposed by the wavelets in this algorithm,while the fusion rule with the maximum absolute value coefficient and based on the local maximum average value variance is used to fuse the high frequency sub-images.The experimental results show that the algorithm evidently improves inclusive information of image fusion,eliminates the circumstances of local contrast polarity reversal,tones the definition of image fusion and commendably maintains the edge detail of input images.