Multicomponent Driven Consistency Priors for Simultaneous Decomposition and Pansharpening
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing(2019)
摘要
Pansharpening is also known as the fusion of a low-resolution (LR) multispectral (MS) image and a high-resolution (HR) panchromatic (Pan) image of the same scene, which is an effective way to improve the spatial resolution of the LR MS image so as to obtain an HR MS image (i.e., pansharpened MS image). In this article, we propose a novel multicomponent consistency priors driven variational model for simultaneous decomposition and pansharpening (SDP) in a unified optimization framework. Specifically, the proposed SDP model particularly decomposes the Pan and MS images into cartoon, structure and texture components, and fully exploits the multicomponent consistency priors on these cartoon, structure and texture components of the Pan and MS images. Thus, the proposed SDP model can suitably characterize the different properties of these multicomponents so that these multicomponents can be well preserved. Moreover, the proposed SDP model is actually a band-coupled model, which can fully preserve the intrinsic structural correlation among the MS bands, because the MS image is actually a spatial-spectral strongly correlated cube. Then, an efficient iterative algorithm based on the forward-backward splitting technique is designed to solve the proposed SDP model. Finally, we compare the proposed SDP method with some state-of-the-art methods on various satellite datasets, and the experimental results demonstrate the effectiveness of the proposed method, which can perform higher spectral and spatial qualities than the other methods.
更多查看译文
关键词
Spatial resolution,Correlation,Optimization,Remote sensing,Multiresolution analysis,Earth
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络