Spectral image classification method and system based on local information constraint and sparse representation

Abstract

The invention provides a spectral image classification method and system based on local information constraint and sparse representation, and the method comprises the steps: measuring the Euclidean distance between a to-be-classified pixel spectrum and a known label pixel spectrum in a spectral image, and determining the range of a dictionary set according to the size of the distance; establishinga mathematical model of spectral image classification based on the category information to obtain a corresponding optimization problem; and solving the abundance coefficient of the spectrum to be classified according to the optimization problem, and classifying the spectral image according to the solution enabling the objective function to take the minimum value. According to the method, the Euclidean distance is introduced into the optimization model, local constraints are formed for the dictionary set, and the robustness of the algorithm is improved by uniformly considering the Euclidean distance between spectral vectors and correlation differences. In addition, prior information of a spectrum library is utilized in the optimization model, and the accuracy of the optimal solution is ensured. Compared with a traditional method, the technical scheme has the advantages of being small in calculated amount and high in classification accuracy.

Description

technical field [0001] The present invention relates to the field of spectral image classification, in particular, the present invention relates to a technical scheme of spectral image classification based on local information constraints and sparse representation. Background technique [0002] Hyperspectral images can provide rich spatial and spectral information. Each pixel in the spectral image contains spectral data of nearly hundreds of bands in the ultraviolet, visible, near-infrared, mid-infrared and thermal infrared bands. In recent years, material classification has become the most basic technical requirement of spectral images, and has been applied in urban planning, geological exploration, environmental monitoring and other fields. [0003] Among the proposed classification methods, one of the most representative methods is support vector machine (SVM). Even in the case of small samples, SVM can achieve satisfactory classification results [1]. In recent years, s...

AI Technical Summary

Problems solved by technology

[0004] Although the above representation methods based on single-pixel classification have achieved good results in hyperspectral images, they have not simultaneously considered multiple differences between the test data and the labeled training set.

Images