High-spectral image classification method based on discrimination and robust multi-feature extraction

Abstract

The invention provides a high-spectral image classification method based on discrimination and robust multi-feature extraction. Different spatial spectral features in high-spectral images are projected into a common low-rank discrimination subspace, and significance of the different features is remained. The method comprises that the features are assigned with different weights according to the significance; a low-dimensional subspace with robust and discrimination features is learned; the different features are projected to the low-dimension subspace; and a support vector machine is used to classify the new features. The dimensions of the features are reduced, noise robustness is realized, noise interference in the high-spectral image can be defended, the significance of different features can be maintained, the obtained features has discrimination characteristic via exiting labels, and the classification effect is improved effectively.

Description

technical field [0001] The invention belongs to the field of image processing, in particular to a hyperspectral image classification method based on identification and robust multi-feature extraction. Background technique [0002] Hyperspectral remote sensing images have very high spectral resolution, and are increasingly used in geological exploration, earth resource investigation, and precision agriculture. Hyperspectral images have very high spectral dimensions, and directly classifying the pixels of hyperspectral images is prone to Hughes phenomenon. Therefore, the method of feature extraction is applied to hyperspectral images. [0003] With the development of hyperspectral image processing technology, algorithms for extracting different types of hyperspectral image features have also emerged, such as texture features, Gable features, etc. With the help of various features, the classification effect of hyperspectral images can be further improved. However, there are s...

AI Technical Summary

Problems solved by technology

However, there are still some problems in this way: (1) the importance of different features for classification is often different, but this is often ignored; (2) spectral images are usually disturbed by noise, so the use of multiple features will bring more much noise

[0004] For the noise problem, low-rank technology is a very effective anti-noise method. However, the low-rank theory is only used for the restoration of the original data, but not for the dimensionality reduction of the feature. Therefore, the feature dimension of the hyperspectral image is too large. problem, the low-rank method also lacks effective means of dealing with

Images