基于多小波包变换的遥感影像融合

Remote-sensed image fusion based on multiwavelet packet transform

多小波变换以完善的小波基结构,可以在影像处理中获得比单小波变换更精确的处理结果;单小波包变换以精细的频域分割,进一步提高了单小波变换对影像的细节信息处理能力。多小波包变换将两者结合起来,综合了它们的优点,同时又弥补了它们的不足,使其在一些影像处理领域更具优势。在对多小波包变换分析之后,提出了一种基于多小波包变换的遥感影像融合方法,并将该方法与多小波变换融合方法(即基于多小波变换的影像融合方法)进行了对比分析。实验结果表明:由该方法所得的融合影像,无论是在视觉效果上还是在评价指标上都要优于由多小波变换融合方法得到的融合影像。

Compared with wavelet transform,multiwavelet transform gains a more precise performance for it’s better wavelet base,and wavelet packet transform achieves a stronger ability to process high-frequency information for it’s finer space-split capability.Multiwavelet packet transform puts their advantages together,therefore exhibits more satisfactory result.The paper firstly analyses it’s basic theory.Then proposes a remote-sensed image fusion method based on multiwavelet packet transform.Experiment testifies that the method can get a better fusing result than the fusion method based on multiwavelet transform.