Hyperspectral image classification method based on adaptive kernel sparse representation and multiple features

Abstract

The invention discloses a hyperspectral image classification method based on adaptive kernel sparse representation and multiple features, and the method comprises the steps: (1) collecting an originalhyperspectral image, carrying out the extraction of a plurality of features of the original hyperspectral image X0, and obtaining a feature space X formed through combination; (2) randomly selectinga part of pixel points from X as a training sample set D, and taking the remaining pixel points as a test sample set; (3) constructing a basic kernel for each feature, and calculating a composite kernel according to each basic kernel and the weight value thereof; and (4) classifying each test sample x to obtain a classification result. According to the method, a multi-kernel learning method is fused into an adaptive kernel sparse representation classifier, a basic kernel is constructed based on different feature descriptions, and the weight value of the basic kernel is calculated, so that thecorrelation between different features can be utilized, the difference of various features can be reserved, and the classification precision can be improved.

Description

technical field [0001] The invention relates to the technical field of hyperspectral data processing, in particular to a hyperspectral image classification method based on adaptive kernel sparse representation and multi-features, which can be applied to practical engineering fields such as aerospace remote sensing and material detection. Background technique [0002] Hyperspectral image classification technology is a research hotspot in the field of remote sensing. Its goal is to classify each spectral pixel into a specific category based on spectral information and learned spatial information. In order to achieve this goal, many classification methods have been proposed, including SVM, MLR, neural network, adaptive artificial immune network and so on. However, these methods only utilize the spectral information of the hyperspectral image, ignoring its spatial information, thus producing noisy classification results. [0003] Making full use of the spatial information of hy...

AI Technical Summary

Problems solved by technology

[0004] Purpose of the invention: In order to overcome the deficiencies of the prior art, the present invention provides a hyperspectral image classification method based on adaptive kernel sparse representation and multi-features, which solves the problem that the pixels in the multi-feature space tend to be linear and cannot Fully consider the similarity and difference between multiple features, and the problem of poor classification results

Images