Multispectral and Panchromatic Image Fusion Method Based on Dual Sparse Nonnegative Matrix Factorization

Abstract

The invention discloses a multi-spectral and panchromatic image fusion method based on dual sparse non-negative matrix decomposition, which mainly solves the problems of fuzzy spatial information and spectral distortion in the prior art. The steps are: (1) respectively input the low spatial resolution multispectral image, high spatial resolution panchromatic image, spectral degradation matrix, and spatial degradation matrix; (2) respectively input the low spatial resolution multispectral image and high spatial resolution Column vectorization of panchromatic images with spatial resolution in blocks; (3) performing dual sparse non-negative matrix decomposition on the image after column vectorization to obtain a dictionary with high spatial resolution and a coefficient matrix with spectral information; (4) transforming The high spatial resolution dictionary and the coefficient matrix with spectral information are multiplied to obtain a column vectorized high spatial resolution multispectral image, and restore it to a high spatial resolution multispectral image. The invention can obtain accurate spatial and spectral information, and can be used in remote sensing fields such as target identification, ground object classification and environmental monitoring.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to an image fusion method in remote sensing images, which can be used in remote sensing fields such as target recognition, ground object classification and environmental monitoring. Background technique [0002] Due to the physical limitation of optical sensor imaging, remote sensing satellites usually can only obtain multispectral images with low spatial resolution and panchromatic images with high spatial resolution. Multispectral images have rich spectral information, but low spatial resolution; panchromatic images have high spatial resolution, rich spatial information, but less spectral information. In order to obtain multispectral images with high spatial resolution, people use image fusion technology to comprehensively utilize the spatial information of panchromatic images and the spectral information of multispectral images, and finally obtain more accurate and...

AI Technical Summary

Problems solved by technology

Although this method can better improve the spatial resolution of hyperspectral images, it is mainly aimed at the fusion of hyperspectral images and visible light images, and cannot achieve the fusion of multispectral images and panchromatic images. The fusion of spectral image and panchromatic image will produce serious spectral distortion

Images