基于双树紧支剪切波变换的遥感图像融合

Remote Image Fusion Based on Dual Tree Compactly Supported Shearlet Transform

摘要: 提出一种基于双树紧支剪切波的遥感图像融合算法。紧支剪切波变换是剪切波理论的空域实现,因其步骤中包含了传统的离散小波变换(DWT),引入了移变性。分析移变性对图像融合的影响,采用双树复数小波抑制紧支剪切波的移变性,与主分量分析变换(PCA)或IHS变换相结合,提出遥感图像融合算法。通过QuickBird和IKONOS的三组数据的进行实验,结果表明提出的融合方法性能优于基于DWT、à trous小波、Curvelet以及基于频域实现剪切波(à trous shearlet)方法。

Abstract: In this paper, a remote image fusion method based on dual tree compactly supported shearlet transform (DT CSST) is proposed. In the method, the shift variant property is compensated by the dual tree structure, DT CSST combining with the PCA or IHS transform is applied in the remote image fusion. Through the QuickBird and IKONOS data, the experiments suggest that the proposed method is superior to other methods based on other multi-scale transforms, such as DWT, dual tree complex wavelet transform (DT CWT), à trous wavelet, curvelet and à trous shearlet transform.