Remote sensing image reconstruction method based on reference image structure constraint and non-convex low rank constraint

Abstract

The invention discloses a remote sensing image reconstruction method based on a reference image structure constraint and a non-convex low rank constraint. The method comprises the following steps: structural similarity constraints for a target image and a reference image after high-order filter are firstly built; then, the non-convex low rank constraint is used for replacing a compressed sensing L1norm for constraining a target image sparse coefficient, and a remote sensing image sparse optimization reconstruction model is built and solved. The method of the invention has the beneficial effects that the high-order structure feature vector of the reference image serves as a prior constraint, a generalized non-convex low rank kernel norm serves as a target image sparse coefficient constraint, the complementary advantages of the two are used for constructing an image reconstruction model, and the reconstruction precision of the target image is improved.

Description

technical field [0001] The invention relates to a signal reconstruction method for multi-source remote sensing image data, in particular to a remote sensing image reconstruction method based on reference image structure constraints and non-convex low-rank constraints. Background technique [0002] In the field of remote sensing, the same area usually contains multi-source and multi-temporal images, and these remote sensing images have different spectral characteristics, temporal resolution and spatial resolution. In some remote sensing applications, we only have some observation images, and we cannot obtain the original images of a certain area at a certain time. If we need the information of these original images, we need to reconstruct the remote sensing images through reconstruction methods. Redundant information on spectral and temporal scales of remote sensing images corresponding to spatial locations is helpful for reconstruction. Different sensors carried in the same...

AI Technical Summary

Problems solved by technology

[0004] The main disadvantages of simply injecting the structural information of remote sensing images into the reconstructed image as reference constraint information are: the injection process is indiscriminate, unable to accurately describe the local structure of the spectral image, ignoring the correlation of sparse coefficients, and unable to describe the local sparseness of the image degree, the introduction of Gibbs effect and other issues

[0005] For the problems in related technologies, no effective solutions have been proposed yet

Images