Restoration Of Hyperspectral Images Based On Low-rank Prior Of Tensors

Hyperspectral remote sensing data carries a large amount of space and spectral information,which creates conditions for people to study the characteristics of surface objects and conduct ground object identification,and makes it widely used and concerned in many fields.However,the acquisition process of hyperspectral remote sensing data is affected by various factors,and there are various degrees of degradation,which will affect the subsequent processing and application.Therefore,the restoration of hyperspectral images is an important part of digital image processing,which is of great significance to the improvement of image quality information,which creates conditions for people to study the characteristics of surface objects and conduct ground object identification,and makes it widely used and concerned in many fields.However,the acquisition process of hyperspectral remote sensing data is affected by various factors,and there are various degrees of degradation,which will affect the subsequent processing and application.Therefore,the restoration of hyperspectral images is an important part of digital image processing,which is of great significance to the improvement of image quality.On the basis of summarizing the current situation of hyperspectral restoration,this paper mines the sparse and low-rank prior knowledge of hyperspectral data through in-depth analysis of the characteristics of hyperspectral data,studies the hyperspectral classification and target detection technology of combined space spectral information,and designs the corresponding efficient algorithm.The main work and research results of this paper are as follows:The development status of hyperspectral image processing and analysis and some problems that need to be solved are expounded,the existing theories and methods of hyperspectral image analysis are summarized,and the application potential and existing problems in hyperspectral image processing and analysis based on models such as low-rank and tensor space are analyzed in detail.For example: LRTV model,tensor recovery model based on kronecker-basis-representation.Objective with the wealth of available spatial and spectral information,hyperspectral remote sensing images have been used for many remote sensing applications and have attracted considerable attention.However,most hyperspectral remote sensing images suffer from degradation because of the limitation of electronic devices,the influence of poor illumination,and the distortion of atmospheric transmission.The degraded data can lead to seriously inaccurate results in the subsequent processing.Thus,on the basis of the low rank constraint and the total variation regularization in this study,a new model is proposed to realize hyperspectral remote sensing images super-resolution reconstruction.First,two types of low rank based priori information in the hyperspectral remote sensing images,i.e.,the low rank-based priori information in the spectral domain and the low rank-based priori information in the spatial domain,are explored.Then,on the basis of the low rank-based priori information in the spectral domain,a low-rank constraint model based on tensor truncated nuclear norm is proposed to achieve more accurate approximation of the rank function;Subsequently,on the basis of the low rank-based priori information in the spatial domain,a total variation regularization is proposed to keep sharp edges and more detailed information of the original image.Finally,the low-rank constraint model based on tensor truncated nuclear norm and total variation models are integrated.As such,the new restoration model based on the low-rank constraint and total variation regularization possesses the advantage of the two aforementioned models.This study tests the performance of the proposed method with a set of challenging hyperspectral remote sensing images.The super-resolution reconstruction results of the proposed method are compared with those of several related methods.At the same time,the peak signal-to-noise ratio and structural similarity indices are adopted to provide quantitative assessments of the results of the experiments.All the experiments prove that the proposed model achieves better visual quality and quantitative indices than those of several existing related methods.The proposed model relies on the low rank-based priori knowledge in the spectral and spatial domains,could effectively achieve super-resolution reconstruction of hyperspectral remote sensing images after being blurred and down-sampled.Finally,the proposed model can be extended to other fields of remote sensing applications.