Hyperspectral image panchromatic sharpening method based on spectral dimension control convolutional neural network

Abstract

The invention provides a hyperspectral image panchromatic sharpening method based on a spectral dimension control convolutional neural network. The hyperspectral image panchromatic sharpening method comprises the following steps: reading original hyperspectral image data; synthesizing corresponding panchromatic image data by using the hyperspectral image; preprocessing the image data to obtain a training sample; constructing a spectral dimension control convolutional neural network structure; inputting the training sample into a spectral dimension control convolutional neural network, and stabilizing a loss function at a minimum value by using an adaptive moment estimation algorithm, thereby obtaining an optimal network structure parameter; inputting the test samples subjected to the samepreprocessing into an optimal spectral dimension control convolutional neural network structure, and outputting and processing to obtain a high-resolution hyperspectral image. According to the method,the characteristics of the hyperspectral image are combined, the missing detail information of the hyperspectral image can be better restored by controlling the spectral dimension of the hyperspectral image, and the spectral distortion phenomenon of a sharpening result is reduced.

Description

technical field [0001] The invention relates to the field of remote sensing images, in particular to a hyperspectral image panchromatic sharpening method based on a spectral dimension control convolutional neural network. Background technique [0002] Hyperspectral image processing technology is a research hotspot in the field of remote sensing and image processing. Compared with multispectral images and panchromatic images, hyperspectral images contain very rich spectral information in addition to spatial information, which has great application value in many aspects. However, an obvious shortcoming of hyperspectral imagery is its low spatial resolution. In order to overcome this shortcoming, a basic strategy that can be adopted is to fuse the hyperspectral image with its registered high-spatial-resolution panchromatic image to obtain a high-spatial-resolution hyperspectral image. This fusion process is usually referred to as It is called hyperspectral image panchromatic ...

AI Technical Summary

Problems solved by technology

[0004] In view of this, in order to solve the above-mentioned problems in the prior art, the present invention provides a hyperspectral image panchromatic sharpening method based on spectral dimension control convolutional neural network , combined with the characteristics of the hyperspectral image, through the control of the spectral dimension of the hyperspectral image, it can better repair the missing detail information of the hyperspectral image and reduce the spectral distortion of the sharpening result

Images