Hyper-spectral remote sensing image classification method based on self-adaptive hierarchical multi-scale

Abstract

The invention discloses a hyper-spectral remote sensing image classification method based on self-adaptive hierarchical multi-scale. The method comprises the following steps that step 1, the irregular neighborhood structure of pixels is calculated according to the spectral angle; step 2, the scale parameters of each layer are determined according to the Ka measure hierarchy in the irregular neighborhood structure, the corresponding kernel matrix of each layer is calculated layer by layer and the weight of the kernel function of each layer is obtained by using the maximum projection variance so that a self-adaptive hierarchical multi-scale kernel function is obtained; and step 3, a hyper-spectral image is mapped to the kernel space of the self-adaptive hierarchical multi-scale kernel function obtained in the step 2, and the pixels to be measured are linearly represented by a dictionary formed on the basis of the known training sample pixels so that a reconstruction sparse matrix is obtained and the pixels to be measured are allocated to the optimal reconstruction category. The hyper-spectral remote sensing data can be rapidly and accurately classified.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing, in particular to a hyperspectral remote sensing image classification method based on self-adaptive hierarchical multi-scale. Background technique [0002] Hyperspectral image classification is an important application direction of hyperspectral image remote sensing. Hyperspectral image classification determines a category mark for each pixel, which is an analysis method to describe the types of ground objects. The result of the category of ground objects can clearly reflect the spatial distribution of ground objects, which is convenient for people to understand and discover laws. Compared with traditional remote sensing image classification, hyperspectral image classification has the following difficulties: 1) high data dimension, insufficient training samples; 2) many bands, high correlation between bands; 3) obvious intra-class differences; 4) large amount of data , oft...

AI Technical Summary

Problems solved by technology

Although the sparse representation classification method has a good classification effect, it is still difficult to distinguish those linearly inseparable data. In order to overcome this defect, some scholars have adopted the kernel sparse representation algorithm, which inherits the excellent performance of the kernel method. The data is projected into a high-dimensional feature space, so that the data is linearly separable

[0004] However, there are still problems with the above methods: 1) Due to the high correlation of dictionaries in the basic framework of sparse representation, the hyperspectral classification problem based on sparse representation is ill-posed, and there are a large number of local approximate solutions. Prior information introduces regular terms to constrain the range of solutions

However, there is no kernel function that is universally applicable to all classification situations

And a single kernel function limits the possibility of mining more hyperspectral information, especially for multi-category situations

Moreover, a single kernel function has no advantage in maintaining the classification accuracy and the generalization ability of the model.

3) The problem of parameter selection in the kernel method is also a big problem. The traditional method of parameter testing is cross-checking, but while the parameters increase, it is necessary to train the data repeatedly, and the time cost is large.

Images