High-spectral image pseudo color visualization method based on joint information

Abstract

The invention discloses a high-spectral image pseudo color visualization method based on joint information and belongs to the remote sensing image processing field. The method comprises the followingsteps: 1) calculating covariance matrixes of all wavebands of a high-spectral image; 2) among a high-spectral waveband set, finding all combinations of three wavebands; 3) for each combination, extracting, from the covariance matrixes, corresponding elements to form a new covariance matrix; 4) for the covariance matrix of each combination, calculating joint information; 5) searching fora waveband combination with the maximum joint information, the combination being a high-spectral pseudo color visualized synthesis waveband combination; and 6) calculating variance values of all wavebands of the optimal waveband combination, and performing order descending in accordance with the variance values, thecorresponding visualized synthesis waveband sequentially being a red waveband, a green waveband and a blue waveband. Through definition of joint combination of the high-spectral image, unified description of information content and correlation of high-spectral wave bands is realized and the high-spectral image pseudo color visualization quality is improved.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing, and in particular relates to a hyperspectral image three-band synthesis pseudo-color visualization method based on joint information. Background technique [0002] The development of hyperspectral remote sensing technology provides an important basis for the fine identification of ground objects. Compared with ordinary remote sensing images, an important advantage of hyperspectral images is the addition of rich spectral dimension information, which can realize continuous spectral sampling from visible light to infrared. The rich spectral information of hyperspectral images can fully reflect the differences in the physical structure characteristics and chemical composition characteristics of the target, so as to realize the fine distinction of the spectral dimension of ground objects. [0003] While hyperspectral images have hundreds of bands that bring rich spectral infor...

AI Technical Summary

Problems solved by technology

[0006] Some band selection methods consider the correlation of the band when designing the criterion function. For example, OIF uses the ratio of the variance of the band and the correlation coefficient as the criterion function, but when the number of bands is large, the computational complexity is extremely high. It is the amount of information in the band, and the correlation between the bands is insufficiently suppressed, so it is not suitable for pseudo-color visualization of hyperspectral images

Images