基于有限脊波变换的图像融合研究  

Image Fusion Based on Finite Ridgelet Transform

作　　者：刘坤[1] 郭雷[1] 李晖晖[1] 常威威[1] 

机构地区：[1]西北工业大学自动化学院,陕西西安710072

出　　处：《光电子．激光》2007年第11期1382-1385,共4页Journal of Optoelectronics·Laser

基　　金：国家自然科学基金资助项目(60175001)

摘　　要：将有限脊波变换(FRIT)应用于图像融合能更好地提取图像边缘特征,为融合提取更多的特征信息。基于FRIT的图像融合算法是将图像进行FRIT,在不同的频率域利用融合规则融合有限脊波系数,最后通过重构得到融合图像。通过对多聚焦图像的实验,对比了基于FRIT的图像融合算法和基于小波变换的融合算法,并对实验结果进行了主观和客观的评价,试验结果表明,基于FRIT的图像融合算法取得了比基于小波变换算法更好的融合结果,是一种有效可行的图像融合算法。Finite ridgelet transform（FRIT） is applied to image fusion,the character of original images is effectively extracted, and more important information is preserved. The fusion method based on FRIT needs two steps. Firstly,the original images are decomposed using finite ridgelet transform, then the coefficients are fused with the fusion regular in the different frequency bands, and finally the fused coefficients are reconstructed to obtain fusion results. Multi-focus images are taken as experimental data,subjective and objective standard are used to evaluate the results,and the comparison with results based on wavelet transform shows that this method gets much better fusion results tan wavelet. And image fusion algorithm based on FRIT is effective and feasible.

关 键 词：图像融合 有限脊波变换(FRIT) 有限Radon变换(FRAT) 脊波变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]