Hyperspectral dimensionality reduction method based on spectral-spatial information maintenance

Abstract

The present invention discloses a hyperspectral dimensionality reduction method based on spectral-spatial information maintenance.<{EN3}>The method introduces waveband selection into the hyperspectral dimension reduction and employs the waveband selection to select a waveband subset having the maximum difference; and the method fully utilizes the spectral-spatial similarity between samples on the selected waveband subset so as to improve the ground object recognition rate, and can be used for fields such as remote sensing ground object ground object observation, military reconnaissance and criminology assistance, etc. The method introduces waveband selection into the hyperspectral dimension reduction and employs the waveband selection to select a waveband subset having the maximum difference; and the method fully utilizes the spectral-spatial similarity between samples on the selected waveband subset so as to improve the ground object recognition rate, and can be used for fields such as remote sensing ground object ground object observation, military reconnaissance and criminology assistance, etc.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to a hyperspectral dimension reduction method. Background technique [0002] With the development of remote sensing technology and imaging spectrometers, the application of hyperspectral remote sensing images is becoming more and more extensive. However, it has the characteristics of large number of bands and huge amount of data, which brings great difficulties to the classification and recognition of hyperspectral images, such as high information redundancy, large space required for data storage, long processing time, and due to hyperspectral The image has a large number of bands, which is prone to the curse of dimensionality, that is, the classification accuracy decreases. Therefore, in the case of ensuring the classification and recognition rate of ground objects, it is very necessary to reduce the amount of data and save resources. Hyperspectral dimensionality reducti...

AI Technical Summary

Problems solved by technology

[0007] The present invention proposes a hyperspectral dimensionality reduction method based on spatial spectrum information preservation, which mainly aims at the high redundancy existing in the original hyperspectral data, improves data expressivity, and overcomes the problem of poor discrimination of existing methods

Images