Tensorized embedded hyperspectral image classification method based on sparse low-rank regular graph

Abstract

The invention discloses a tensorized embedded hyperspectral image classification method based on a sparse low-rank regular graph. The method overcomes the shortcomings in the prior art that classification is performed without fully utilizing hyperspectral image sample information and neighborhood information. The method comprises steps of: (1) inputting a hyperspectral image; (2) determining a training sample set and a test sample set; (3) constructing an adjacent matrix of the training sample set; (4) constructing a sparse low-rank regular graph; (5) determining an airspace training sample set and an airspace test sample set; (6) constructing a tensorized graph embedding model; (7) performing dimension reduction on the airspace test sample set; (8) performing hyperspectral image classification; and (9) outputting classified images. The method has an accurate classification effect at the edges and the homogeneous regions of the hyperspectral images and can be used for classifying the hyperspectral images.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a hyperspectral image classification method based on quantitative embedding of sparse low-rank regular graphs in the technical field of hyperspectral image classification. The invention can be used to classify the hyperspectral images. Background technique [0002] The improvement of spatial domain and spectral domain resolution of hyperspectral images provides more abundant information for classification, but also brings great challenges. Traditional classification methods, including decision tree classification, artificial neural network classification and support vector machine classification, only classify features from the spectral domain level. However, hyperspectral remote sensing data not only contains rich spectral information of surface objects, but also has specific description and expression of surface object characteristics in two different dimensions...

AI Technical Summary

Problems solved by technology

Although this method can make good use of the category information of ground objects, it still has the disadvantage that the large number of labeled samples required for training the model will increase the difficulty of manually labeling samples, and the insufficient utilization of unlabeled samples will It causes a waste of spectral information, and the insufficient utilization of neighborhood information in the sample space will seriously affect the classification effect of images in edge and homogeneous regions.

Although this method can quickly classify hyperspectral images, it still has the disadvantage that the undirected graph obtained through sparse representation cannot well maintain the characteristics of the global data structure, resulting in large errors in classification results.

Images