Block sparse representation mode and structure dictionary-based hyperspectral image compression method

Abstract

The invention relates to a block sparse representation mode and structure dictionary-based hyperspectral image compression method. The method is realized under the premise that signals with similar structures have similar sparse representation modes, and after the signals with similar structures undergo structured sparse representation, sparse coefficients express similar block sparse representation modes, namely, non-zero elements present to same positions. The method comprises the following steps of: at an onboard encoding end, carrying out sparse representation on an image under a structuredictionary; carrying out compressed encoding on a sparse coefficient on the basis of a sparse representation mode; and finally transmitting encoded information to a decoding end. According to the method, high-complexity nonlinear reconfiguration does not need to be carried out at the decoding end, the sparse coefficient can be recovered according to transfer coefficient label information, position information and mean value information, and the original image can be reestablished through multiplying the sparse coefficient by the dictionary, so that the reestablishing time is remarkably shortened.

Description

technical field [0001] The invention belongs to the field of hyperspectral remote sensing image compression, and relates to a hyperspectral image compression method based on a block sparse expression mode and a structure dictionary. Background technique [0002] Compared with natural images, hyperspectral images contain two-dimensional spatial information and one-dimensional spectral information. Wherein, each spectral band corresponds to a two-dimensional image, and pixels at the same position in different bands form a spectral curve. Taking advantage of the differences in spectral curves of different ground features, hyperspectral images are widely used in various fields of the national economy. With the popularization of the application of high-resolution remote sensing imaging technology, how to effectively compress the challenges of the surge in transmission and storage data due to the significant improvement in the spectral and spatial resolution of remote sensing ima...

AI Technical Summary

Problems solved by technology

[0006] 1. The decoding end uses a nonlinear reconstruction method with high complexity, which has high computational complexity and low efficiency;

[0007] 2. The structural characteristics of the hyperspectral image are not fully utilized, and the unstructured dictionary is used to sparsely represent the image, and the structural similarity of the reconstructed image is low;

[0008] 3. The similar block sparse expression mode between sparse coefficients is ignored, the sparse coefficients are not encoded, and the overall image compression rate is not high enough

Images