Spectral image optimal waveband selection method and system based on improved OIF, and storage medium

Abstract

The invention discloses a spectral image optimal waveband selection method and system based on an improved OIF, and a storage medium. According to the invention, the Optimal Index Factor (OIF) is improved based on image information entropy, the method can be used for hyperspectral band selection and can also be used for multispectral and hyperspectral band selection, the selected band is the optimal combined band, and the number of the selected bands is set by a user himself/herself. The OIF is improved on the basis of the image information entropy to be used for selecting the optimal combinedwave band, the correlation information of spectral images among different wave bands and information of the information amount contained in the spectral images are fully utilized, the correlation ofthe selected combined wave band is small, the contained information amount is large, and redundant information is little.

Description

Technical field[0001]The present invention relates to the field of spectral image band selection, and more particularly to a selection method, system, and storage medium based on the optimal wave band based on improved OIF.Background technique[0002]The spectral imaging technology is a combination of spectral techniques and imaging techniques, on the basis of ordinary two-dimensional space imaging, increases one-dimensional spectrum information. The spectral imaging technique can acquire a number of narrow and spectral continuous image data from the ultraviolet bands of the electromagnetic band to the medium infrared regions, depending on the spectral resolution of the imaging spectrometer, can be divided into multi-spectral images, high spectrum images, and ultra-spectrum images. In general, the wave band of the multi-spectral imager is generally from several tens of the wave bands of high spectral imaging, and the imaging band of the super-spectroscopy is tens of thousands or even ...

AI Technical Summary

Problems solved by technology

At present, there are only some patent reports on hyperspectral band selection. For example, 201911232855.6 proposes a hyperspectral image band selection method based on deep learning. The deep learning model is difficult to train and requires high computer performance. It is used to contain hundreds of bands. The hyperspectral image data selection process is time-consuming and the model is complex; 201911027959.3 realizes hyperspectral optimal band selection based on standard deviation and correlation. Compared with the deep learning method, the method of this patent is relatively simple and fast, but the method is essentially Optimum Index Factor (OIF), proposed by the American scholar Chavez, uses the standard deviation to describe the amount of information of the spectral image, which has limitations

In addition, the above-mentioned invention is only for hyperspectral, and does not involve multispectral and hyperspectral

Images