CN102322957A - Spectrum drifting detection method for interference type hyperspectral imager - Google Patents

Abstract

The invention discloses a spectrum drifting detection method for an interference type hyperspectral imager. The method comprises the following steps: (1) converting the spectroscopic data of a hyperspectral imager into radiance on entrance pupil position of each wave band to serve as spectrum radiance to be detected; (2) obtaining the radiance on entrance pupil position of satellite height; (3) obtaining radiance on entrance pupil position of each wave band by the response function of the hyperspectral imager to serve as reference spectrum radiance; (4) comparing the positions of the spectrum radiance to be detected and the reference spectrum radiance on an atmospheric absorption peak, determining whether the offset of the spectrum radiance to be detected and the reference spectrum radiance by a judgment function; turning to the step (5) if yes; otherwise, regulating the regulation amount of wave band and bandwidth in the response function of the hyperspectral imager, and circularly judging from step (3); and (5) carrying out inverse fast Fourier transform to the interference data of the hyperspectral imager for the hyperspectral imager spectrum drifting amount determined when a constraint condition is satisfied to obtain a spectrum data graph of the hyperspectral imager, which removes spectrum drifting.

Description

A kind of spectral drift detection method of interfere type hyperspectral imager

Technical field

The present invention relates to a kind of spectral drift detection method of interfere type hyperspectral imager, particularly a kind of spectral drift detection method of environment mitigation satellite hyperspectral imager.

Background technology

The interfere type hyperspectral imager is the corresponding relation that utilizes between interferogram and the light source light spectrogram; Through being carried out the inverse fourier transform inverting, interferogram obtains spectrogram; It obtains abundant atural object spatial information, radiation information and spectral information with nano level spectral resolution; Be widely used and the military and civilian field, have great using value and vast potential for future development.

Wave numbers such as the normal employing of interfere type hyperspectral imager are divided into a plurality of passages with spectroscopic data at interval; Reach the passage more than 100; The bandwidth of each passage has only several nanometers; Therefore the spectral calibration precision not only influences the precision of radiation calibration, and the quantification that also directly influences the hyperspectral imager data is used.

Hyperspectral imager can carry out spectrum to imager through laboratory calibration and the calibration of star polishing wax; But because imager changes in the rapid variation of space environments such as emission process or aloft high vibration, othermohygrometer; Cause the performance variation of image device; Cause laboratory calibration and the calibration of star polishing wax to have a certain distance, thereby to uncertainty and error that quantitative remote sensing causes, the quantification that influences remotely-sensed data is used with actual conditions.Therefore need be through carrying out the detection of spectral drift, to guarantee the application of follow-up quantification at the rail spectral calibration.

Since interfere type hyperspectral imager image-forming principle singularity---what imager received is the Fourier transform information of terrain object spectrum; External only have the powerful moonlet of the U.S. to adopt this image-forming principle; Domestic research and the application that has just begun the interfere type hyperspectral imager; Up to the present, do not see that as yet relevant interfere type hyperspectral imager is in the detection method of rail spectral drift in the domestic and foreign literature data.

Summary of the invention

Technology of the present invention is dealt with problems and is: overcome the deficiency of prior art, a kind of spectral drift detection method of interfere type hyperspectral imager is provided.

Technical solution of the present invention is: a kind of spectral drift detection method of interfere type hyperspectral imager, and step is following:

(1) reads hyperspectral imager interference data figure, this hyperspectral imager interference data is carried out inverse discrete Fourier transformer inverse-discrete obtain hyperspectral imager spectroscopic data figure; Through the absolute radiometric calibration coefficient, this hyperspectral imager spectroscopic data is converted into each wave band entrance pupil place spoke brightness of hyperspectral imager, with it as the spectral radiance of testing to be checked;

Ground actual measurement face of land parameter and atmospheric parameter during (2) according to satellite imagery utilize the radiation delivery model to obtain the entrance pupil place spoke brightness of satellite altitude;

(3) obtain each wave band entrance pupil place spoke brightness of hyperspectral imager through the hyperspectral imager response function, with its spectral radiance as a reference;

(4) comparison position of testing spectral radiance and the brightness of reference spectra spoke at the Atmospheric Absorption peak to be checked during satellite imagery; Confirm through discriminant function whether the side-play amount of testing spectral radiance and the brightness of reference spectra spoke to be checked reaches constraint condition, if reach then change step (5); Otherwise the adjustment amount of adjustment hyperspectral imager response function medium wavelength and bandwidth is from step (3) beginning cycle criterion;

(5) the hyperspectral imager spectral drift amount of confirming in the time of will satisfying constraint condition is carried out inverse fourier transform again to the hyperspectral imager interference data and is obtained the spectroscopic data figure after hyperspectral imager is removed spectral drift.

Discriminant function in the said step (4) is following:

min?g＝min{(1-κ)σ+κα}

Wherein: κ is the weight coefficient ratio, κ ∈ [0,1];

N is the sum of hyperspectral imager wave band;

$\mathrm{\σ}=\sqrt{\frac{1}{n-1}\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{(L_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)-L_{\mathrm{referrence}}\left(\mathrm{\λ}\right))}^2}$

$\mathrm{\α}={\cos }^{-1}\left(\frac{\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{L}_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)L_{\mathrm{referrence}}\left(\mathrm{\λ}\right)}{\sqrt{\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{\left(L_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)\right)}^2\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{\left(L_{\mathrm{referrence}}\left(\mathrm{\λ}\right)\right)}^2}/\frac{\mathrm{\π}}{2}}\right)$

L _Unchecked(λ) the entrance pupil place spoke brightness of corresponding optometry spectrum wave band to be checked

L _Referrence(λ) the entrance pupil place spoke brightness of corresponding reference spectra wave band;

λ is the wavelength of hyperspectral imager wave band.

The present invention compared with prior art beneficial effect is:

(1) the present invention is from the application aspect of interfere type hyperspectral imager data, and the intrinsic absorption peak characteristic of the atmosphere when utilizing satellite imagery has realized that the interfere type hyperspectral imager does not have the technical matters that detects at the rail spectral drift at present.

(2) the present invention starts with from the processing of interfere type hyperspectral imager data; After the spectral drift amount is confirmed; Directly feed back to hyperspectral imager data processing source; Again carry out the production of hyperspectral imager spectroscopic data, this makes that the spectral calibration of hyperspectral imager interference data is more pointed and more accurate to follow-up quantification application.

(3) the present invention is according to the imaging characteristic of interfere type hyperspectral imager; Simulate through the large number of ground measured test; Consider many-sided susceptibility such as spectral characteristic matching degree and error irrelevance to the spectral drift amount; Establishing the adjustment amount that adopts with hyperspectral imager response function medium wavelength and bandwidth is the discriminant function of stealthy variable, effectively judges the drifting state of interfere type hyperspectral imager.

Description of drawings

Fig. 1 is a process flow diagram of the present invention.

Embodiment

Introduce implementation procedure of the present invention in detail below in conjunction with accompanying drawing 1, concrete steps are following:

(1) reads hyperspectral imager interference data figure, this hyperspectral imager interference data is carried out inverse fourier transform obtain hyperspectral imager spectroscopic data figure; Through the absolute radiometric calibration coefficient, this hyperspectral imager spectroscopic data is converted into each wave band entrance pupil place spoke brightness of hyperspectral imager, with it as the spectral radiance of testing to be checked;

(1.1) hyperspectral imager spectroscopic data

Read the hyperspectral imager interference data,, 1. calculate the hyperspectral imager spectroscopic data according to formula according to laboratory spectral calibration result:

B(λ)＝Λ·I(Δ) ①

In the formula: Δ is the optical path difference of coherent light beam;

λ is the wavelength of hyperspectral imager wave band;

The corresponding hyperspectral imager interference data of I (Δ) intensity;

The corresponding hyperspectral imager spectroscopic data of B (λ) intensity;

The corresponding inverse fourier transform matrix of coefficients of Λ.

(1.2) spectral radiance of testing to be checked

With above-mentioned hyperspectral imager spectroscopic data, each wave band entrance pupil place spoke brightness of 2. calculating hyperspectral imager according to formula, with it as the spectral radiance of testing to be checked:

L _unchecked(λ)＝k(λ)B(λ)+L ₀(λ) ②

Wherein:

L _Unchecked(λ) the entrance pupil place spoke brightness of corresponding optometry spectrum wave band to be checked

K (λ), L ₀(λ) calibration coefficient of corresponding hyperspectral imager wave band can obtain from the accompanying document of hyperspectral imager data.

Ground actual measurement face of land parameter and atmospheric parameter during (2) according to satellite imagery; Utilize the radiation delivery model to obtain the entrance pupil place spoke brightness of satellite altitude; Obtain each wave band entrance pupil place spoke brightness of hyperspectral imager through the hyperspectral imager response function, with its spectral radiance as a reference;

(2.1) the entrance pupil place spoke brightness of satellite altitude

Through atmospheric radiation transmission such as MODTRAN, HITRAN or SCIATRAN etc., face of land parameter, atmospheric parameter and other parameter input models of actual measurement can obtain the entrance pupil place spoke brightness of satellite altitude when satellite was passed by;

(2.2) reference spectra spoke brightness

3. the satellite altitude entrance pupil place spoke brightness that aforementioned calculation is obtained calculate each wave band entrance pupil place spoke brightness of hyperspectral imager according to formula, with its spectral radiance as a reference:

$L_{\mathrm{referrence}}\left(\mathrm{\λ}\right)=\frac{\∫f\left(\mathrm{\λ}\right)S\left(\mathrm{\λ}\right)\mathrm{d\λ}}{\∫f\left(\mathrm{\λ}\right)\mathrm{d\λ}}$ ③

Wherein: L _Referrence(λ) the entrance pupil place spoke brightness of corresponding reference spectra wave band;

The entrance pupil place spoke brightness of the corresponding satellite altitude of S (λ);

The spectral response functions of the corresponding hyperspectral imager of f (λ);

(2.3) spectral response functions of hyperspectral imager

The spectral response functions of hyperspectral imager, adopt Gaussian function 4. to simulate:

$f\left(\mathrm{\λ}\right)=e^{-{\left[\frac{\mathrm{\λ}-({\mathrm{\λ}}_c+\mathrm{\Δ\λ})}{(\mathrm{FWTH}+\mathrm{\ΔFWTH})/\left(2\sqrt{2\ln 2}\right)}\right]}^2}$ ④

Wherein: λ _cCentre wavelength for the hyperspectral imager wave band;

FWTH is the bandwidth of hyperspectral imager wave band;

Δ λ, Δ FWTH are the adjustment amount of wavelength and bandwidth;

(3) comparison position of testing spectral radiance and the brightness of reference spectra spoke at the Atmospheric Absorption peak to be checked during satellite imagery; Confirm through discriminant function whether the side-play amount of testing spectral radiance and the brightness of reference spectra spoke to be checked reaches constraint condition, if reach then change step (4); Otherwise the adjustment amount of adjustment hyperspectral imager response function medium wavelength and bandwidth is from step (2.2) beginning cycle criterion;

(3.1) Atmospheric Absorption peak

In the remote sensing satellite imaging process, Atmospheric Absorption gas plays attenuation to emittance, and differing absorption gas is different to the absorption of wavelength, and it has tangible wavelength selectivity, like H ₂O has four strong absorption bandses, O at 400～1000nm ₂At 400～1000nm three strong absorption bandses are arranged; These absorption bandses are inherent features of atmosphere, utilize its absorption peak characteristic can realize that the spectral drift of hyperspectral imager detects.

(3.2) discriminant function

Adopt discriminant function 5.; Through continuous adjustment input parameter (Δ λ, Δ FWTH), make reference spectra spoke brightness and to be checked test spectral radiance and carry out the optimization coupling, when discriminant function reaches hour (being above-mentioned constraint condition); Stop to judge output hyperspectral imager spectral drift amount;

min?g＝min{(1-κ)σ+κα} ⑤

Wherein: κ is the weight coefficient ratio, κ ∈ [0,1];

N is the sum of hyperspectral imager wave band;

$\mathrm{\σ}=\sqrt{\frac{1}{n-1}\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{(L_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)-L_{\mathrm{refeerrence}}\left(\mathrm{\λ}\right))}^2}$

$\mathrm{\α}={\cos }^{-1}\left(\frac{\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{L}_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)L_{\mathrm{referrence}}\left(\mathrm{\λ}\right)}{\sqrt{\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{\left(L_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)\right)}^2\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{\left(L_{\mathrm{referrence}}\left(\mathrm{\λ}\right)\right)}^2}/\frac{\mathrm{\π}}{2}}\right)$

(4) the hyperspectral imager spectral drift amount of confirming in the time of will satisfying constraint condition is carried out inverse fourier transform again to the hyperspectral imager interference data and is obtained the spectroscopic data figure after hyperspectral imager is removed spectral drift.

The unspecified part of the present invention belongs to general knowledge as well known to those skilled in the art.

Claims (2)

1. the spectral drift detection method of an interfere type hyperspectral imager is characterized in that step is following:

(1) reads hyperspectral imager interference data figure, this hyperspectral imager interference data is carried out inverse discrete Fourier transformer inverse-discrete obtain hyperspectral imager spectroscopic data figure; Through the absolute radiometric calibration coefficient, this hyperspectral imager spectroscopic data is converted into each wave band entrance pupil place spoke brightness of hyperspectral imager, with it as the spectral radiance of testing to be checked;

Ground actual measurement face of land parameter and atmospheric parameter during (2) according to satellite imagery utilize the radiation delivery model to obtain the entrance pupil place spoke brightness of satellite altitude;

(3) obtain each wave band entrance pupil place spoke brightness of hyperspectral imager through the hyperspectral imager response function, with its spectral radiance as a reference;

(4) comparison position of testing spectral radiance and the brightness of reference spectra spoke at the Atmospheric Absorption peak to be checked during satellite imagery; Confirm through discriminant function whether the side-play amount of testing spectral radiance and the brightness of reference spectra spoke to be checked reaches constraint condition, if reach then change step (5); Otherwise the adjustment amount of adjustment hyperspectral imager response function medium wavelength and bandwidth is from step (3) beginning cycle criterion;

(5) the hyperspectral imager spectral drift amount of confirming in the time of will satisfying constraint condition is carried out inverse fourier transform again to the hyperspectral imager interference data and is obtained the spectroscopic data figure after hyperspectral imager is removed spectral drift.

2. the spectral drift detection method of a kind of interfere type hyperspectral imager according to claim 1 is characterized in that the discriminant function in the said step (4) is following:

ming＝min{(1-κ)σ+κα}

Wherein: κ is the weight coefficient ratio, κ ∈ [0,1];

N is the sum of hyperspectral imager wave band;

$\mathrm{\σ}=\sqrt{\frac{1}{n-1}\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{(L_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)-L_{\mathrm{referrence}}\left(\mathrm{\λ}\right))}^2}$

$\mathrm{\α}={\cos }^{-1}\left(\frac{\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{L}_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)L_{\mathrm{referrence}}\left(\mathrm{\λ}\right)}{\sqrt{\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{\left(L_{\mathrm{unchecked}}\left(\mathrm{\λ}\right)\right)}^2\underset{\mathrm{\λ}=1}{\overset{n}{\mathrm{\Σ}}}{\left(L_{\mathrm{referrence}}\left(\mathrm{\λ}\right)\right)}^2}/\frac{\mathrm{\π}}{2}}\right)$

L _Unchecked(λ) the entrance pupil place spoke brightness of corresponding optometry spectrum wave band to be checked

L _Referrence(λ) the entrance pupil place spoke brightness of corresponding reference spectra wave band;

λ is the wavelength of hyperspectral imager wave band.

Images