CN-115790845-B - Method, system and medium for correcting spectrum bending

CN115790845BCN 115790845 BCN115790845 BCN 115790845BCN-115790845-B

Abstract

The invention provides a method, a system and a medium for correcting spectrum bending, and relates to the technical field of a spectrum imager and satellite remote sensing, wherein the method comprises the following steps that S1, a solar spectrum image is selected as sample data for correcting spectrum bending, and curve fitting is carried out on the sample data; and step S3, correcting each column of the image to be corrected by utilizing the time shifting property of Fourier transformation according to the curvature of each column, and completing correction of all columns of data to obtain corrected complete spectrum image data. The invention can process the image data by utilizing the time shift characteristic of discrete Fourier transformation without increasing the design and manufacturing difficulty of a light-splitting component or a slit and complex calculation, and simply and intuitively complete the spectrum bending correction.

Inventors

GU QIANG
WANG XIYANG
ZHANG JIAN
YUAN YUAN
XU ZHI
LI ZHEN
LI CHUAN

Assignees

上海卫星工程研究所

Dates

Publication Date: 20260508
Application Date: 20221125

Claims (9)

1. A method of spectral curvature correction, comprising: S1, selecting a solar spectrum image as sample data for spectrum bending correction, and performing curve fitting on the sample data; S2, calculating the curvature of each column of the image to be corrected according to the fitted curve; and step S3, correcting each column of the image to be corrected by utilizing the time shifting property of Fourier transformation according to the curvature of each column, and finishing the correction of all column data to obtain corrected complete spectrum image data.
2. The method of spectral curvature correction according to claim 1, wherein the step S1 comprises: S1.1, selecting a solar spectrum image as sample data for spectrum bending correction, reading the image, and storing the image into a two-dimensional array S (M, N), wherein M is more than or equal to 1 and less than or equal to M, N is more than or equal to 1 and less than or equal to N, M is a space dimension, and N is a spectrum dimension; S1.2, finding out the line number of the minimum value of each column from S (m, n), and storing the line number of the found minimum value into a one-dimensional array y (m); And step S1.3, performing quadratic polynomial fitting on y (m) to obtain coefficients a, b and c of y (m) =am 2 +bm+c.
3. The method of spectral curvature correction according to claim 1, wherein step S2 comprises: Step S2.1, calculating the trough value of the fitting curve according to the fitted quadratic polynomial, namely The value at miny, i.e S2.2, reading a solar spectrum image to be corrected, storing the solar spectrum image into a two-dimensional array q (M, N), and establishing a two-bit array p (M, N) for storing data after spectrum bending correction, wherein M is more than or equal to 1 and less than or equal to M, N is more than or equal to 1 and less than or equal to N, M is a space dimension, and N is a spectrum dimension; s2.3, reading the ith column (i is more than or equal to 1 and less than or equal to M) of q (M, n), and storing the ith column into a one-dimensional array x i (n); And S2.4, calculating a difference delta y i ＝(ai 2 +bi+c) -miny of the ith column relative to the valley value of the fitting curve according to the fitting quadratic polynomial.
4. A method of spectral curvature correction according to claim 3, wherein step S3 comprises: S3.1, performing discrete Fourier transform on X i (N) to obtain X i (k) (k is more than or equal to 1 and less than or equal to N), namely X i (k)＝fft(x i (N)); Step S3.2, X i (k) is compared with the time shift property of the discrete Fourier transform Multiplying and inverse Fourier transforming and taking the real part to obtain the corrected data x i ' (n), i.e Step S3.3, saving the corrected i-th column data to the i-th column of p (m, n), i.e., p (i, n) =x i ' (n); and repeating the steps S2.3 to S3.3, and finishing the correction of all the q (m, n) column data to obtain corrected complete spectrum image data p (m, n).
5. A system for spectral curvature correction, comprising: The module M1 is used for selecting a solar spectrum image as sample data for spectrum bending correction and performing curve fitting on the sample data; The module M2 calculates the curvature of each column of the image to be corrected according to the fitted curve; And the module M3 is used for correcting each column of the image to be corrected by utilizing the time shifting property of Fourier transformation according to the curvature of each column, so that the correction of all column data is completed, and corrected complete spectrum image data is obtained.
6. The system for spectral curvature correction according to claim 5, wherein the module M1 comprises: selecting a solar spectrum image as sample data for spectrum bending correction, reading the image, and storing the image into a two-dimensional array s (M, N), wherein M is more than or equal to 1 and less than or equal to M, N is more than or equal to 1 and less than or equal to N, M is a space dimension, and N is a spectrum dimension; The module M1.2 is used for finding out the line number of the minimum value of each column from s (M, n), and storing the line number of the found minimum value into a one-dimensional array y (M); And (3) performing quadratic polynomial fitting on y (M) to obtain coefficients a, b and c of y (M) =am 2 +bm+c.
7. The system for spectral curvature correction according to claim 5, wherein the module M2 comprises: module M2.1. Based on the fitted quadratic polynomial, the trough value of the fitted curve is calculated, i.e The value at miny, i.e The module M2.2 is used for reading a solar spectrum image to be corrected, storing the solar spectrum image into a two-dimensional array q (M, N), and establishing a two-bit array p (M, N) for storing data after spectrum bending correction, wherein M is more than or equal to 1 and less than or equal to M, N is more than or equal to 1 and less than or equal to N, M is a space dimension, and N is a spectrum dimension; Reading the ith column (i is more than or equal to 1 and less than or equal to M) of q (M, n) and storing the ith column into a one-dimensional array x i (n); module M2.4 calculates the difference deltay i ＝(ai 2 +bi+c) -miny between the ith column and the trough of the fitted curve from the fitted quadratic polynomial.
8. The system for spectral curvature correction according to claim 7, wherein the module M3 comprises: Performing discrete Fourier transform on X i (N) to obtain X i (k) (k is more than or equal to 1 and less than or equal to N), namely X i (k)＝fft(x i (N)); module M3.2 taking advantage of the time-shifted nature of the discrete Fourier transform, X i (k) is compared with Multiplying and inverse Fourier transforming and taking the real part to obtain the corrected data x i ' (n), i.e The module M3.3 saves the corrected ith column data to the ith column of p (M, n), i.e. p (i, n) =x i ' (n); and repeating the modules M2.3 to M3.3, and finishing the correction of all q (M, n) column data to obtain corrected complete spectrum image data p (M, n).
9. A computer-readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the steps of the method of any one of claims 1 to 4.

Description

Method, system and medium for correcting spectrum bending Technical Field The invention relates to the technical field of spectrum imagers and satellite remote sensing, in particular to a mathematical method for correcting the bending of a spectrum image, and particularly relates to a method for curve fitting of the bending of the spectrum and correcting the bending of the spectrum by utilizing the time shifting property of discrete Fourier transformation. And more particularly to a method, system, and medium for spectral curvature correction. Background The broom pushing scanning type spectrum imager uses the area array detector to scan and image the two-dimensional view field in a solid self-scanning mode. However, the light rays at different positions in the length direction of the slit pass through different refraction surfaces of the light splitting element, so that the spectrum of the space dimension is bent. Spectral warping causes images to produce spectral aliasing. Therefore, the correction of the spectrum bending is necessary, and the method has higher practical application value. Currently, optical methods and mathematical methods are mainly used for correcting the spectrum bending. The optical method mainly improves the spectrum bending effect by optimizing the design of the light-splitting component or the slit. As disclosed in the patent publication No. CN104034419a, a combination of a prism and a grating is used as a spectroscopic device, and the characteristic that the grating and the prism are curved in opposite directions of spectral lines (the grating spectral line is curved in the wavelength direction and the prism spectral line is curved in the short wavelength direction) is utilized to correct the spectral line curvature of the center wavelength so that the spectral line curvature of other wavelengths is symmetrically distributed about the center wavelength. However, the structure has different degrees of line bending in different wavebands due to the inconsistent line bending produced by the grating and the prism in the full operating band. The invention patent with publication number CN108663119A describes a method for correcting spectral line bending of a spectrometer by designing a bending slit and compensating the spectral line bending brought by the whole system by using the bending entrance slit. However, the machining of curved slits is difficult, and it is difficult to manufacture acceptable slits. Mathematical methods such as "hyperspectral imager image spectral curvature correction method study-chestnut-doctor paper" introduce a method of interpolation to eliminate spectral aliasing by resampling the radiant energy data collected by the system to non-offset wavelength scaling data to correspond the corrected output image to the input uncorrected image. The specific interpolation method can adopt linear interpolation, piecewise four-point cubic Lagrange interpolation and cubic spline interpolation. The essence of interpolation is to estimate the data of an unknown point by knowing the value of the sample point, requiring the data to be corrected to be convolved with a finite length continuous impulse response called the interpolation kernel. The method involves convolution operation, is large in calculation amount, and is not visual in bending correction process and is obscure to understand. The invention patent publication No. CN104316183A describes a method for computing the weight contribution of a plurality of points closest to a non-off-center wavelength by computing the distance between the points and the non-off-center wavelength, and recalculating the assignment of the spectral radiant energy corresponding to the center wavelength according to the weight contribution value. The method considers the correlation between adjacent wave bands, has high spectrum fidelity and strong universality, but the weight value calculation and distribution process is complex, and the calculated amount is large. Disclosure of Invention Aiming at the defects in the prior art, the invention provides a method, a system and a medium for correcting spectrum bending. According to the method, the system and the medium for correcting the spectrum bending, which are provided by the invention, the scheme is as follows: In a first aspect, there is provided a method of spectral bow correction, the method comprising: S1, selecting a solar spectrum image as sample data for spectrum bending correction, and performing curve fitting on the sample data; S2, calculating the curvature of each column of the image to be corrected according to the fitted curve; and step S3, correcting each column of the image to be corrected by utilizing the time shifting property of Fourier transformation according to the curvature of each column, and finishing the correction of all column data to obtain corrected complete spectrum image data. Preferably, the step S1 includes: S1.1, selecting a solar spectrum image as sampl