Hyperspectral image band selection method and system and electronic equipment

Abstract

The invention provides a hyperspectral image waveband selection method and system and electronic equipment, and the method comprises the steps: S100, obtaining a hyperspectral image cube, and segmenting the hyperspectral image cube into a plurality of regions through PCA and a superpixel segmentation algorithm; s200, learning low-dimensional potential features corresponding to the regions by adopting a layering strategy, obtaining potential feature matrixes of the regions, extracting average potential features of the hyperspectral image cube, and constructing an average potential feature matrix; s300, fusing the potential feature matrix of each region with the average potential feature matrix to obtain a low-dimensional self-characterization matrix of the hyperspectral image cube; and S400, carrying out clustering on the low-dimensional self-characterization matrix to obtain an optimal wave band combination. The method has the beneficial effect that the accuracy of hyperspectral image waveband selection is effectively improved.

Description

technical field [0001] The invention relates to the technical field of computer science, and in particular, to a method, system and electronic device for selecting a hyperspectral image band. Background technique [0002] Hyperspectral sensors capture spectral and spatial information of target scenes by collecting dozens or even hundreds of consecutive hyperspectral bands. Compared with RGB images, hyperspectral images have richer information and higher resolution. With the continuous development and maturity of hyperspectral imaging technology and image classification technology, hyperspectral imaging is widely used in various fields, such as salient target detection, medical image processing and mineral exploration. However, hyperspectral images also have some problems. For example, within a certain range, the spectra of adjacent bands are very similar, which increases the computational complexity of image classification and affects the final classification accuracy. In ...

AI Technical Summary

Problems solved by technology

[0003] At present, the existing technology mainly uses unsupervised band selection and supervised band selection to achieve dimensionality reduction processing of hyperspectral images. Although it shows good dimensionality reduction performance, there are also shortcomings that cannot be ignored, resulting in inaccurate band selection.

For example, in the feature representation process of the band, it is easy to cause the loss of spatial structure information; the original high-dimensional pixel features used often contain noise, which affects the band selection performance; when using the spatial structure information, the weight difference between different regions is ignored. sex

Images