Method for removing strip noises of imaging spectrometer

Abstract

The invention discloses a method for removing strip noises of an imaging spectrometer. The method comprises the steps of performing two-dimensional discrete wavelet decomposition on a to-be-denoised waveband image for certain times in combination with a decomposition formula by using a wavelet basis function Haar; selecting an image with strip noises meeting the requirement at a long waveband as areference image, and performing the two-dimensional discrete wavelet decomposition on the reference image for the same times as the to-be-denoised waveband image; performing histogram matching on vertical detail components obtained after the two-dimensional discrete wavelet decomposition of the reference image, and replacing vertical detail components obtained after the two-dimensional discrete wavelet decomposition of the to-be-denoised waveband image with the same dimension; and recombining vertical detail component replacement results through inverse wavelet transform to finish the de-noising. The method can reserve original information of the image to the maximum extent while eliminating the strip noises.

Description

technical field [0001] The invention relates to the technical field of spectral image processing, in particular to a method for removing band noise of an imaging spectrometer. Background technique [0002] Imaging spectrometer was developed on the basis of multispectral remote sensing imaging technology in the 1980s. It acquires hyperspectral images of scenery or targets with high spectral resolution. information, to achieve the effect of "map integration". Due to the huge advantage of the high spectral resolution of the imaging spectrometer, it has become a hot spot in the field of remote sensing to obtain many continuous-band spectral images of ground objects while observing the earth in space, so as to achieve the purpose of directly identifying the materials on the earth's surface from space. The main technical means of space observation of the Earth. [0003] Band noise is a kind of band noise that often appears in hyperspectral image data, and generally has the chara...

AI Technical Summary

Problems solved by technology

The disadvantages of the above scheme are: 1. This scheme is only applicable to the hyperspectral image before geometric correction, and the ground object information of the image is required to be relatively simple, otherwise the stripping effect will be poor

2. If the image is small or the ground object information is complex and the gray level distribution is uneven, using the moment matching method will destroy the spectral information of the original image, and will produce a "band effect" along the direction perpendicular to the CCD scanning, resulting in image distortion

However, the disadvantage of this scheme is that it will lose a large amount of non-noise image information while removing bands, reduce image clarity, and change image spectral characteristics.

Images