CN102749138B - Spectrum calibration method based on sun and atmosphere characteristic spectrum in hyperspectral remote sensor flight - Google Patents

Abstract

The invention discloses a spectrum calibration method in hyperspectral remote sensor flight. The spectrum calibration method includes that a sun fraunhofer line and an atmosphere characteristic absorbing line are used as a standard reference spectrum under the normal working condition in the hyperspectral remote sensor flight, and determination of center wavelength offset and spectral bandwidth is achieved through matching of equivalent spectrum reflection rate rho*(lambda) and equivalent spectrum reflection rate of lowpass filtering. The spectrum calibration process does not depend on measurement of surface feature spectrum reflection rate, and effects of indeterminacy of radiation calibration of a remote sensor are reduced. The spectrum calibration method is achieved through a look-up table method and improves executive efficiency of a program. The spectrum calibration method can achieve spectrum calibration during onboard or satellite-borne hyperspectral remote sensor flight and spectrum calibration of a field surface feature spectrum radiometer smaller than 0.2 nm and has significant meaning in hyperspectral remote sensing.

Description

High-spectrum remote-sensing device is the spectral calibration method based on the sun and Atmospheric Characteristics spectral line in-flight

Technical field

The invention belongs to remote optical sensing scientific domain, relate to a kind of high-spectrum remote-sensing device spectral calibration method based on the sun and Atmospheric Characteristics spectral line in-flight.

Background technology

High-spectrum remote-sensing device can obtain the meticulous continuous spectrum information of target, and it had brought into play more and more important effect in fields such as national defense and military, survey of territorial resources, precision agriculture, environmental monitoring, atmospheric explorations in recent years.Take and determine that the spectral calibration that high-spectrum remote-sensing device spectrum channel centre wavelength and bandwidth are object is the prerequisite of its telemetry quantification application.The calibration of pre-flight Laboratory Spectra is generally to scan or the method for characteristic absorption wavelength plate Spectral matching realizes by monochromator, but because of reasons such as atmospheric pressure, temperature variation and high vibrations, centre wavelength and the bandwidth of high-spectrum remote-sensing device likely change under remote sensor state of flight.Research shows, very little spectral calibration results change can be brought very large measuring error, as can produce at strong water vapor absorption wave band ± 25% spoke brightness measurement error of the center wavelength shift of the remote sensor 1nm of 10nm bandwidth, therefore, the aloft spectral calibration of high-spectrum remote-sensing device is one of requisite link in its data quantitative process.

The gaseous absorption lines such as the sun fraunhofer line in atmosphere profile and oxygen, carbon dioxide, steam are from the absorption spectrum of atom or molecule, wherein sun fraunhofer line is very stable line-spectra, and in atmosphere, the line style of gas molecule absorption spectrum has close relationship with temperature pressure.As far back as last century, just by scientist, be used for realizing the centre wavelength of spectrometer is calibrated, in recent years, the people such as Geffen utilized the outer solar spectrum of the atmosphere of superelevation spectral resolution to realize the spectral calibration to spaceborne high spectral radiometers such as GOME, SCIAMACHY.According to the literature, last century the nineties, the people such as the raiser Goetz of imaging spectrometer concept and Green be take at first Reflectivity for Growing Season and the atmospheric parameter of actual measurement and are input data, utilize radiation transmitting software Modtran to generate simulated data, and adopt Spectral matching method to realize the spectral calibration of airborne hyperspectral remote sensor AVIRIS.Subsequently, the people such as Gao are not in the situation that needing earth's surface actual measurement reflectivity, utilize Spectral matching algorithm to obtain Hyperion, AVIRIS, the sensors such as PHILLS are in the side-play amount of Atmospheric Absorption band center wavelength, the people such as German scholar Guanter in 2006 utilize telluric lines by the method that combines with atmospheric correction to imaging spectrometer ROSIS, CHRIS, Hymap, AVIRIS etc. have carried out spectral calibration test, obtained the skew of centre wavelength, the people such as Barry of U.S. TRW Ltd. (US) One Space Park, Redondo Beach CA 90278 U.S.A. utilize telluric lines to obtain the remote sensor center wavelength shift in process in orbit by the method that makes Hyperion face limit observation.The centre wavelength that the people such as France scholar Didier Ramon utilize atmospheric pressure inverting to realize this remote sensor in conjunction with the method for oxygen 760nm Absorption Line by the special track of design MERIS is calibrated.

Summary of the invention

The object of this invention is to provide a kind of high-spectrum remote-sensing device spectral calibration method in-flight, take sun fraunhofer line and Atmospheric Characteristics Absorption Line as standard reference, by equivalent spectrum reflectivity ρ ^*(λ) realize determining of center wavelength shift amount and spectral bandwidth with the mating of equivalent spectrum reflectivity through low-pass filtering.

The technical solution used in the present invention is:

High-spectrum remote-sensing device is the spectral calibration method based on the sun and Atmospheric Characteristics spectral line in-flight, it is characterized in that, concrete steps are as follows:

The spoke brightness at remote sensor entrance pupil place is mainly comprised of atmospheric path radiation, target reflection, the three part combined actions of background radiation around;

The spoke brightness L of remote sensor entrance pupil _tOA(λ) 0 use formula can be expressed as:

$L_{\mathrm{TOA}}\left(\mathrm{\λ}\right)=L_0\left(\mathrm{\λ}\right)+L_{a-g}\left(\mathrm{\λ}\right)+\frac{E_g\left(\mathrm{\λ}\right)}{\mathrm{\π}}\mathrm{\ρ}\left(\mathrm{\λ}\right)T_{\↑}\left(\mathrm{\λ}\right)---\left(1\right)$

Wherein, L ₀(λ) be the spectral radiance that atmospheric path radiation produces, E _g(λ) be that the sun is incident on tellurian built-up radiation, ρ (λ) is the spectral reflectivity of atural object, T _↑(λ) be remote sensor-destination path transmitance, L _a-g(λ) be the spectral radiance that background radiation around produces;

Therefore, clutter reflections rate is expressed as:

$\mathrm{\ρ}\left(\mathrm{\λ}\right)=\frac{L_{\mathrm{TOA}}\left(\mathrm{\λ}\right)\·\mathrm{\π}}{E_g\left(\mathrm{\λ}\right)\·T_{\↑}\left(\mathrm{\λ}\right)}-\frac{(L_0\left(\mathrm{\λ}\right)+L_{a-g}\left(\mathrm{\λ}\right))\·\mathrm{\π}}{E_g\left(\mathrm{\λ}\right)\·T_{\↑}\left(\mathrm{\λ}\right)}---\left(2\right)$

Order ${\mathrm{\ρ}}^{*}\left(\mathrm{\λ}\right)=\frac{L_{\mathrm{TOA}}\left(\mathrm{\λ}\right)\·\mathrm{\π}}{E_g\left(\mathrm{\λ}\right)\·T_{\↑}\left(\mathrm{\λ}\right)}$ ， ${\mathrm{\ρ}}^{\mathrm{atm}}\left(\mathrm{\λ}\right)=\frac{(L_0\left(\mathrm{\λ}\right)+L_{a-g}\left(\mathrm{\λ}\right))\·\mathrm{\π}}{E_g\left(\mathrm{\λ}\right)\·T_{\↑}\left(\mathrm{\λ}\right)}$ ；

Therefore, 2 formulas can be expressed as:

$\mathrm{\ρ}\left(\mathrm{\λ}\right)={\mathrm{\ρ}}^{*}\left(\mathrm{\λ}\right)-{\mathrm{\ρ}}^{\mathrm{atm}}\left(\mathrm{\λ}\right)---\left(3\right)$

Wherein, ρ ^atm(λ) only relevant with atmospheric condition and surrounding environment, irrelevant with remote sensor characteristic, therefore in calculating, spectral calibration optimization do not consider this;

When the centre wavelength of remote sensor or bandwidth change, ρ ^*(λ) at Atmospheric Absorption wave band, produce sharp keen projection and depression;

The objective function of spectral calibration optimized algorithm is:

${\mathrm{\χ}}^2({\mathrm{\δ}}_1,{\mathrm{\δ}}_2)=\underset{i=n_1}{\overset{n_2}{\mathrm{\Σ}}}{[{\mathrm{\ρ}}_i^{*}({\mathrm{\δ}}_1,{\mathrm{\δ}}_2)-{\mathrm{\ρ}}_i^{\mathrm{smooth}}]}^2---\left(4\right)$

${\mathrm{\ρ}}_i^{\mathrm{smooth}}=\frac{1}{N}\underset{j=i-2}{\overset{i+2}{\mathrm{\Σ}}}{\mathrm{\ρ}}_j({\mathrm{\δ}}_1,{\mathrm{\δ}}_2)---\left(5\right)$

Wherein, the ρ of i wave band after low-pass filtering ^*(λ), δ ₁δ ₂respectively the side-play amount of centre wavelength and spectral bandwidth, n ₁, n ₂it is near the spectral band number gas absorption wave band of choosing;

The method that adopts look-up table (HITRAN database) is calculated in the optimization of spectral calibration, searches and makes (4) formula obtain minimum value respectively in center wavelength shift and spectral bandwidth variation range, has realized determining of center wavelength shift amount and spectral bandwidth.

Principle of the present invention is: equivalent spectrum reflectivity ρ ^*(λ) with through the mating of the equivalent spectrum reflectivity of low-pass filtering, the situation when more approaching remote sensor spectrum channel characteristic and do not change through the equivalent spectrum reflectivity of low-pass filtering.Constantly change the side-play amount of centre wavelength and the size of spectral bandwidth, by least square method, obtain best matching result, the matching algorithm of spectrum adopts the mode of look-up table.Owing to being coupling between spectral reflectivity, therefore eliminated the impact of ground spectrum albedo measurement in the spectral radiance match spectrum calibrating method of documents and materials reports.

Advantage of the present invention is:

(1) spectral calibration process of the present invention does not rely on the measurement of reflectance, has reduced the impact that remote sensor radiation calibration uncertainty is brought;

(2) spectral calibration algorithm of the present invention is realized by the method for look-up table, has improved the execution efficiency of program;

(3) the present invention can realize airborne or spaceborne high-spectrum remote-sensing device in-flight and field ground feature spectral radiometer is less than the spectral calibration of 0.2nm, has important meaning in high-spectrum remote-sensing.

Accompanying drawing explanation

Fig. 1 is solar radiation-ground-remote sensor interaction schematic diagram.

Fig. 2 is the search procedure example of spectral calibration algorithm.

Embodiment

As shown in Figure 1, be interactional schematic diagram between solar radiation-ground-remote sensor.High-spectrum remote-sensing device is the spectral calibration method based on the sun and Atmospheric Characteristics spectral line in-flight, and concrete steps are as follows:

The spoke brightness at remote sensor entrance pupil place is mainly comprised of atmospheric path radiation, target reflection, the three part combined actions of background radiation around;

The spoke brightness L of remote sensor entrance pupil _tOA(λ) with formula, can be expressed as:

$L_{\mathrm{TOA}}\left(\mathrm{\λ}\right)=L_0\left(\mathrm{\λ}\right)+L_{a-g}\left(\mathrm{\λ}\right)+\frac{E_g\left(\mathrm{\λ}\right)}{\mathrm{\π}}\mathrm{\ρ}\left(\mathrm{\λ}\right)T_{\↑}\left(\mathrm{\λ}\right)---\left(1\right)$

Wherein, L ₀(λ) be the spectral radiance that atmospheric path radiation produces;

E _g(λ) be that the sun is incident on tellurian built-up radiation;

ρ (λ) is the spectral reflectivity of atural object;

T _↑(λ) be remote sensor-destination path transmitance;

T _a-g(λ) be the spectral radiance that background radiation around produces;

Therefore, clutter reflections rate can be expressed as:

$\mathrm{\ρ}\left(\mathrm{\λ}\right)=\frac{L_{\mathrm{TOA}}\left(\mathrm{\λ}\right)\·\mathrm{\π}}{E_g\left(\mathrm{\λ}\right)\·T_{\↑}\left(\mathrm{\π}\right)}-\frac{(L_0\left(\mathrm{\λ}\right)+L_{a-g}\left(\mathrm{\λ}\right))\·\mathrm{\π}}{E_g\left(\mathrm{\λ}\right)\·T_{\↑}\left(\mathrm{\λ}\right)}---\left(2\right)$

Order ${\mathrm{\ρ}}^{*}\left(\mathrm{\λ}\right)=\frac{L_{\mathrm{TOA}}\left(\mathrm{\λ}\right)\·\mathrm{\π}}{E_g\left(\mathrm{\λ}\right)\·T_{\↑}\left(\mathrm{\λ}\right)}$ ， ${\mathrm{\ρ}}^{\mathrm{atm}}\left(\mathrm{\λ}\right)=\frac{(L_0\left(\mathrm{\λ}\right)+L_{a-g}\left(\mathrm{\λ}\right))\·\mathrm{\π}}{E_g\left(\mathrm{\λ}\right)\·T_{\↑}\left(\mathrm{\λ}\right)}$

Therefore 2 formulas can be expressed as:

$\mathrm{\ρ}\left(\mathrm{\λ}\right)={\mathrm{\ρ}}^{*}\left(\mathrm{\λ}\right)-{\mathrm{\ρ}}^{\mathrm{atm}}\left(\mathrm{\λ}\right)---\left(3\right)$

Wherein, ρ ^atm(λ) only relevant with atmospheric condition and surrounding environment, irrelevant with remote sensor characteristic, therefore in calculating, spectral calibration optimization do not consider this;

When the centre wavelength of remote sensor or bandwidth change, ρ ^*(λ) at Atmospheric Absorption wave band, produce sharp keen projection and depression;

The objective function of spectral calibration optimized algorithm is:

${\mathrm{\χ}}^2({\mathrm{\δ}}_1,{\mathrm{\δ}}_2)=\underset{i=n_1}{\overset{n_2}{\mathrm{\Σ}}}{[{\mathrm{\ρ}}_i^{*}({\mathrm{\δ}}_1,{\mathrm{\δ}}_2)-{\mathrm{\ρ}}_i^{\mathrm{smooth}}]}^2---\left(4\right)$

${\mathrm{\ρ}}_i^{\mathrm{smooth}}=\frac{1}{N}\underset{j=i-2}{\overset{i+2}{\mathrm{\Σ}}}{\mathrm{\ρ}}_j({\mathrm{\δ}}_1,{\mathrm{\δ}}_2)---\left(5\right)$

Wherein,

the ρ of i wave band after low-pass filtering ^*(λ), δ ₁δ ₂respectively the side-play amount of centre wavelength and bandwidth, n ₁, n ₂it is near the spectral band number gas absorption wave band of choosing; The method that adopts look-up table (HITRAN database) is calculated in the optimization of spectral calibration, searches respectively and make (4) formula obtain minimum value in the variation range of center wavelength shift and spectral bandwidth, has realized determining of center wavelength shift amount and spectral bandwidth.

Fig. 2 is exemplified as spectral calibration computation process, for highlighting, gets lg (1/ χ ²) be z axle, the skew of centre wavelength is x axle, bandwidth be changed to y axle.

Claims (1)

1. the high-spectrum remote-sensing device spectral calibration method based on the sun and Atmospheric Characteristics spectral line in-flight, the spoke brightness at remote sensor entrance pupil place is mainly comprised of atmospheric path radiation, target reflection, the three part combined actions of background radiation around; It is characterized in that, concrete steps are as follows:

The spoke brightness of remote sensor entrance pupil

with formula, can be expressed as:

Wherein, the spectral radiance that atmospheric path radiation produces,

that the sun is incident on tellurian built-up radiation, the spectral reflectivity of atural object,

remote sensor-destination path transmitance,

it is the spectral radiance that background radiation around produces;

Therefore, clutter reflections rate is expressed as:

Order

Therefore, (2) formula can be expressed as:

Wherein,

only relevant with atmospheric condition and surrounding environment, irrelevant with remote sensor characteristic, therefore in calculating, spectral calibration optimization do not consider this;

When the centre wavelength of remote sensor or bandwidth change,

at Atmospheric Absorption wave band, produce sharp keen projection and depression;

The objective function of spectral calibration optimized algorithm is:

Wherein,

that i wave band is after low-pass filtering

, δ ₁, δ ₂respectively the side-play amount of centre wavelength and spectral bandwidth, n ₁, n ₂it is near the spectral band number gas absorption wave band of choosing;

The method that adopts look-up table is calculated in the optimization of spectral calibration, searches and makes (4) formula obtain minimum value respectively in center wavelength shift and spectral bandwidth variation range, has determined center wavelength shift amount and spectral bandwidth.

Images