Hyperspectral image classification method based on novel neighborhood selection constraints

Abstract

The invention discloses a hyperspectral image classification method based on novel neighborhood selection constraints. The hyperspectral image classification method comprises the following steps: S1,reading hyperspectral image data; S2, selecting a training sample according to the hyperspectral data, and constructing a dictionary D; S3, determining THE adaptive neighborhood set A of a to-be-processed test sample; S4, determining the multi-scale square neighborhood set B of the to-be-processed test sample; S5, determining the final neighborhood set C of the to-be-processed test sample, and constructing a final neighborhood test set J corresponding to the test sample; S6, solving a corresponding coefficient matrix set [psi]; S7, reconstructing the sample, and calculating a reconstructionresidual set R corresponding to the final neighborhoods of different scales; and S8, finally determining the category of the hyperspectral pixel by reconstructing the residual set R. According to themethod, information of neighborhoods of different scales is comprehensively utilized and considered; meanwhile, through the effective constraint of the adaptive neighborhood, the difference of different test pixels is considered, the interference of non-similar neighborhood pixels is avoided respectively, and the method has the advantages of good classification image visual effect, high classification precision and the like.

Description

Technical field [0001] The invention relates to the technical field of remote sensing information processing, in particular to a hyperspectral image classification method based on a novel neighborhood selection constraint. Background technique [0002] Hyperspectral images are collected by optical sensors. Generally, hyperspectral images have the characteristics of high dimensionality and massive information. Hyperspectral image classification is to take all the spectral information contained in each pixel as a whole. The spectral information corresponding to different categories has certain differences due to different electromagnetic energy reflected by ground objects. According to this difference, each image is divided Yuan assigns the sample category label. With the development of technology, collaborative representation classification (CRC) and joint collaborative representation classification (JCRC) have been successfully introduced into hyperspectral image classification,...

AI Technical Summary

Problems solved by technology

[0003] Problems in the CRC classification method: for the pixels in the image, the neighborhood information of the hyperspectral image is not considered

Problems in the JCRC classification method: 1. For different pixels, the neighborhood is a fixed-size square neighborhood, and the neighborhood information is relatively simple and it is easy to introduce interference from different types of pixels

2. For different pixels, the comprehensive neighborhood information in the hyperspectral image has not been adaptively and effectively extracted

[0004] The above problems lead to the inability of the hyperspectral image to be well expressed, resulting in low classification accuracy

Images