CN104361563A - GPS-based (global positioning system based) geometric precision correction method of hyperspectral remote sensing images - Google Patents

Abstract

The invention provides a GPS-based (global positioning system based) geometric precision correction method of hyperspectral remote sensing images. The method includes: reading bil-formatted hyperspectral remote sensing image data acquired by an imaging spectrometer and GPS-recorded attitude information data of an imaging platform; converting longitude and latitude, corresponding to GPS-recorded scanning lines when the imaging spectrometer acquires the hyperspectral remote sensing images, into Gaussian plane rectangular coordinates; correcting images of all frequencies. Geometric distortions of the hyperspectral remote sensing images, generated by unstable movements of the imaging platform, are precisely corrected through the attitude information data of the imaging platform recorded by the high-precision GPS, universality is high, and time and labor consumption is lower; meanwhile, little attitude information data of the imaging platform is required, calculated amount is small, and the need for real-time image correction can be met accordingly.

Description

Based on the high-spectrum remote sensing geometric accurate correction method of GPS

Technical field

The present invention relates to high-spectrum remote sensing geometric correction method, specifically a kind ofly sweep the high-spectrum remote sensing that type imaging spectrometer obtains carry out the method for geometric accurate correction to pushing away.

Background technology

In recent years, in fields such as environmental monitoring, geology, agricultural, medical science and military affairs, traditional panchromatic and coloured image far can not meet the demand of people, and all kinds of spectral imaging technology is applied widely.Imaging spectrometer makes remote sensing technology step into the new stage that simultaneously can obtain earth surface substance spectra information and spatial distribution characteristic information.The sensor of imaging spectrometer detects atural object or target to the reflection of up to a hundred different wave lengths or radiation intensity, forms the spectrum picture be made up of up to a hundred continuous print spectral bands.High-spectrum remote sensing is defined as the 3 D stereo data of two-dimensional space territory and one-dimensional spectrum territory composition, with video sequence image unlike, in high-spectrum remote sensing, each spectral coverage image position is in the spatial domain identical, that is high-spectrum remote sensing is made up of the different spectral coverage image under same field.

Although high-spectrum remote sensing great potential, due to the instability of imaging platform, cause the high-spectrum remote sensing obtained can produce geometric distortion and the application of specific field cannot be met.Therefore, geometry correction is that high spectrum resolution remote sensing technique applies one of problem that must solve.Compared with satellite remote sensing sensor, airborne imaging spectrum instrument has the features such as imaging platform poor stability, flying height be low, the combination of these factors makes the geometric distortion of obtained high-spectrum remote sensing complicated, and the object of geometry correction is exactly eliminate these geometric distortions as far as possible.

The general Method of Remote Sensing Image Geometric Correction used is the polynomial revise method based on ground control point at present, the method utilizes the geometric distortion process of the ground control point information of sufficient amount to remote sensing images to carry out more accurate mathematical simulation, set up the Function Mapping relation of fault image and correcting image, realize the correction projection of fault image.But the impact that the controlled point of the correction accuracy of these class methods is chosen is very large, and precision is not high, and needs artificial calibration and measure a large amount of Ground Control Informations on the spot, takes time and effort.

Summary of the invention

In order to overcome the deficiencies in the prior art, the invention provides a kind of high-spectrum remote sensing geometric accurate correction method based on GPS, object is to overcome the shortcoming that ground control point information that traditional polynomial revise method exists is difficult to obtain and correction accuracy is not high, utilize the synchronous High Precision GPS Data (the attitude information data of record imaging platform) obtained when imaging spectrometer obtains high-spectrum remote sensing, when not needing manual intervention, high-precision geometry correction can be carried out to the high-spectrum remote sensing obtained.

The technical solution adopted for the present invention to solve the technical problems comprises the following steps:

A. data importing;

(A1) the high-spectrum remote sensing data of the bil form that imaging spectrometer obtains are read;

(A2) read the attitude information data of the imaging platform of GPS record, comprise longitude and latitude, angle of pitch ψ, roll angle ω, crab angle κ and flying height H;

(A3) check that the number of scanning lines whether the record number of the attitude information data of GPS record obtains high-spectrum remote sensing during with imaging spectrometer is consistent, if inconsistent, resampling process carried out to the attitude information data of GPS record, mates both making;

(A4) angle of pitch ψ, roll angle ω and crab angle κ in the attitude information data recorded by GPS are converted to Rad;

The longitude and latitude that during the imaging spectrometer acquisition high-spectrum remote sensing B. recorded by GPS, sweep trace is corresponding is converted to Gaussian parabolic line;

(B1) intermediate variable when m article of sweep trace central point longitude and latitude is changed to Gaussian parabolic line during calculating imaging spectrometer acquisition high-spectrum remote sensing, initial value gets m=1; Calculate first intermediate variable α=a × B+b × sin (2B)+c × sin (4B)+d × sin (6B); Wherein, constant coefficient a=6367558.5, b=-16036.48, c=16.828, d=-0.022, B are the latitude of m article of sweep trace central point of GPS record, calculate second intermediate variable β=6399698.902-21562.267 × cos ²b+108.973cos ⁴b-0.612cos ⁶b; Calculate the 3rd intermediate variable η=0.0067385254 × cos ²b; Calculate the 4th intermediate variable L=L ₁-L _c; Wherein, L ₁for the longitude of m article of sweep trace central point of GPS record, L _cfor the median longitudinal of the projection zone that all sweep traces of GPS record are formed;

(B2) Gaussian parabolic line of m article of sweep trace central point in North and South direction during calculating imaging spectrometer acquisition high-spectrum remote sensing $\begin{array}{c}X=\mathrm{\α}+L^2\×\mathrm{\β}\×\sin B\×\cos B/2+L^4\×\mathrm{\β}\×\sin B{\cos }^{3}B\\ \×(5-{\tan }^{2}B+{9\mathrm{\η}}^2+4{\mathrm{\η}}^4)/24+L^6\×\mathrm{\β}\×\sin B{\cos }^{5}B\\ \×(61-58{\tan }^{2}B+{\tan }^{4}B)/720\end{array};$

(B3) Gaussian parabolic line of m article of sweep trace central point on east-west direction during calculating imaging spectrometer acquisition high-spectrum remote sensing $\begin{array}{c}Y=L\×\mathrm{\β}\×\cos B+L^3\×\mathrm{\β}\×{\cos }^{3}B(1-{\tan }^{2}B+{\mathrm{\η}}^2)/6\\ +L^5\×\mathrm{\β}\×{\cos }^{5}B\×(9-18{\tan }^{2}B+{\tan }^{4}B)/120+500000\end{array};$

(B4) the deflection γ=κ+pi/2 of m article of sweep trace during imaging spectrometer acquisition high-spectrum remote sensing is calculated;

(B5) m article of sweep trace central point is calculated when imaging spectrometer obtains high-spectrum remote sensing to the distance D=H/cos ψ of corresponding ground sweep trace central point, wherein, the flying height of m article of sweep trace when H is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record;

(B6) the roll angle ω that during calculating imaging spectrometer acquisition high-spectrum remote sensing, m article of sweep trace i-th pixel is corresponding _i=ω-(N-1) × IFOV/2+i × IFOV, wherein, i is every a line pixel number from left to right, and initial value gets i=1; N is the detection unit number of imaging spectrometer linear array; IFOV is the instantaneous field of view angle of imaging spectrometer;

(B7) m article of sweep trace i-th pixel is calculated when imaging spectrometer obtains high-spectrum remote sensing to the distance S of m article of sweep trace central point _i=D × tan (ω _i) and Gaussian parabolic line component Δ x on north and south and east-west direction _i=S _i× cos (γ) and Δ y _i=S _i× sin (γ);

(B8) the Gaussian parabolic line x that during calculating imaging spectrometer acquisition high-spectrum remote sensing, m article of sweep trace i-th pixel is corresponding _i=X+ Δ x _i, y _i=Y+ Δ y _i, in formula, the Gaussian parabolic line of m article of sweep trace central point on north and south and east-west direction when X, Y are imaging spectrometer acquisition high-spectrum remote sensings;

(B9) repeat (B6) ~ (B8), make i add 1, the Gaussian parabolic line that during calculating imaging spectrometer acquisition high-spectrum remote sensing, on m article of sweep trace, all pixels are corresponding;

(B10) repeat (B1) ~ (B9), make m add 1, the Gaussian parabolic line that during calculating imaging spectrometer acquisition high-spectrum remote sensing, all pixels of all sweep traces are corresponding;

C. image rectification

(C1) pixel resolution is calculated wherein, it is the average flying height of all sweep traces during imaging spectrometer acquisition high-spectrum remote sensing;

(C2) size xsize=(the xmax-xmin)/GR of correcting image in North and South direction is calculated, wherein, when xmax represents that imaging spectrometer obtains high-spectrum remote sensing, Gaussian parabolic line corresponding to all pixels is in the maximal value of North and South direction, and the Gaussian parabolic line that when xmin represents that imaging spectrometer obtains high-spectrum remote sensing, all pixels are corresponding is in the minimum value of North and South direction;

(C3) size ysize=(the ymax-ymin)/GR of correcting image at east-west direction is calculated, wherein, when ymax represents that imaging spectrometer obtains high-spectrum remote sensing, Gaussian parabolic line corresponding to all pixels is in the maximal value of east-west direction, and the Gaussian parabolic line that when ymin represents that imaging spectrometer obtains high-spectrum remote sensing, all pixels are corresponding is in the minimum value of east-west direction;

(C4) calculate line number xindex=(the xmax-x)/GR of high-spectrum remote sensing first pixel correspondence in correcting image that imaging spectrometer obtains, wherein, x is the Gaussian parabolic line of the corresponding North and South direction of first pixel; Circulation performs this step and calculates all pixels line number corresponding in correcting image;

(C5) calculate columns yindex=(the y-ymin)/GR of high-spectrum remote sensing first pixel correspondence in correcting image that imaging spectrometer obtains, wherein, y is the Gaussian parabolic line of the corresponding east-west direction of first pixel; Circulation performs this step and calculates all pixels columns corresponding in correcting image;

(C6) according to line number and the columns of all pixels of high-spectrum remote sensing correspondence in correcting image of imaging spectrometer acquisition, the gray-scale value of all pixels of a kth wave band of the high-spectrum remote sensing obtained by imaging spectrometer is assigned to the corresponding pixel of correcting image, and initial value gets k=1;

(C7) pixel adopting nearest-neighbor method of interpolation to eliminate owing to correcting the inconsistent image correcting data caused of front and back picture size lacks;

(C8) repeat (C6) ~ (C7), make k add 1, complete all band images and correct.

The invention has the beneficial effects as follows: the attitude information data utilizing the imaging platform of high-precision GPS record, fine correction is carried out, highly versatile to the geometric distortion of the high-spectrum remote sensing that imaging platform unsteady motion produces, time saving and energy saving; Meanwhile, because the attitude information data of the imaging platform of the present invention's needs are few, calculated amount is little, therefore, it is possible to meet the needs of real time correction image.

Accompanying drawing explanation

Fig. 1 is the high-spectrum remote-sensing imaging system schematic diagram with imaging spectrometer and high-precision GPS composition;

Fig. 2 is process flow diagram of the present invention;

Fig. 3 is data importing process flow diagram of the present invention;

Fig. 4 is coordinate calculation flow chart of the present invention;

Fig. 5 is image rectification process flow diagram of the present invention.

Embodiment

Below in conjunction with drawings and Examples, the present invention is further described, the present invention includes but be not limited only to following embodiment.

The present invention includes following steps:

A. data importing

(A1) read the high-spectrum remote sensing data of the bil form that imaging spectrometer obtains, data are three-dimensional matrices, and front bidimensional is space dimension, represents spatial information, and the third dimension is spectrum dimension, represents band class information.

(A2) read the attitude information data of the imaging platform of GPS record, comprise longitude and latitude, angle of pitch ψ, roll angle ω, crab angle κ, flying height H.

(A3) check that the number of scanning lines whether the record number of the attitude information data of GPS record obtains high-spectrum remote sensing during with imaging spectrometer is consistent, if inconsistent, resampling process carried out to the attitude information data of GPS record, mates both making.

(A4) angle of pitch ψ, roll angle ω and crab angle κ in the attitude information data recorded by GPS are converted to Rad.

B. coordinate calculates

The longitude and latitude that when what GPS recorded is imaging spectrometer acquisition high-spectrum remote sensing, sweep trace is corresponding, is converted into Gaussian parabolic line.

(B1) intermediate variable when m article of sweep trace central point longitude and latitude is changed to Gaussian parabolic line during calculating imaging spectrometer acquisition high-spectrum remote sensing, initial value gets m=1.The formula calculating first intermediate variable α is: α=a × B+b × sin (2B)+c × sin (4B)+d × sin (6B); Wherein, the constant coefficient calculating α comprises a, b, c and d, a=6367558.5, b=-16036.48, c=16.828, d=-0.022; B is the latitude (radian) of m article of sweep trace central point of GPS record.The formula calculating second intermediate variable β is: β=6399698.902-21562.267 × cos ²b+108.973cos ⁴b-0.612cos ⁶b.The formula calculating the 3rd intermediate variable η is: η=0.0067385254 × cos ²b.The formula calculating the 4th intermediate variable L is: L=L ₁-L _c; Wherein, L ₁for the longitude (radian) of m article of sweep trace central point of GPS record, L _cfor the median longitudinal (radian) of the projection zone that all sweep traces of GPS record are formed.

(B2) formula is utilized $\begin{array}{c}X=\mathrm{\α}+L^2\×\mathrm{\β}\×\sin B\×\cos B/2+L^4\×\mathrm{\β}\×\sin B{\cos }^{3}B\\ \×(5-{\tan }^{2}B+{9\mathrm{\η}}^2+4{\mathrm{\η}}^4)/24+L^6\×\mathrm{\β}\×\sin B{\cos }^{5}B\\ \×(61-58{\tan }^{2}B+{\tan }^{4}B)/720\end{array};$ The m article of Gaussian parabolic line X of sweep trace central point in North and South direction during calculating imaging spectrometer acquisition high-spectrum remote sensing.

(B3) formula is utilized $\begin{array}{c}Y=L\×\mathrm{\β}\×\cos B+L^3\×\mathrm{\β}\×{\cos }^{3}B(1-{\tan }^{2}B+{\mathrm{\η}}^2)/6\\ +L^5\×\mathrm{\β}\×{\cos }^{5}B\×(9-18{\tan }^{2}B+{\tan }^{4}B)/120+500000\end{array};$ The m article of Gaussian parabolic line Y of sweep trace central point on east-west direction during calculating imaging spectrometer acquisition high-spectrum remote sensing.

(B4) the deflection γ of m article of sweep trace during formula γ=κ+pi/2 calculating imaging spectrometer acquisition high-spectrum remote sensing is utilized, wherein, the crab angle of m article of sweep trace when κ is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record.

(B5) formula D=H/cos ψ is utilized to calculate when imaging spectrometer obtains high-spectrum remote sensing m article of sweep trace central point to the distance D of corresponding ground sweep trace central point, wherein, the flying height of m article of sweep trace when H is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record, the angle of pitch of m article of sweep trace when ψ is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record.

(B6) formula ω is utilized _ithe roll angle ω that during=ω-(N-1) × IFOV/2+i × IFOV calculating imaging spectrometer acquisition high-spectrum remote sensing, m article of sweep trace i-th pixel is corresponding _i, wherein, i is every a line pixel number from left to right, and initial value gets i=1; N is the detection unit number of imaging spectrometer linear array, the roll angle that when ω is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record, m article of sweep trace central point is corresponding; IFOV is the instantaneous field of view angle of imaging spectrometer.

(B7) formula S is utilized _i=D × tan (ω _i), Δ x _i=S _i× cos (γ) and Δ y _i=S _i× sin (γ) to calculate when imaging spectrometer obtains high-spectrum remote sensing m article of sweep trace i-th pixel to the distance S of m article of sweep trace central point _iand the Gaussian parabolic line component Δ x on north and south and east-west direction _iwith Δ y _i.

(B8) formula x is utilized _i=X+ Δ x _iand y _i=Y+ Δ y _ithe Gaussian parabolic line x that during calculating imaging spectrometer acquisition high-spectrum remote sensing, m article of sweep trace i-th pixel is corresponding _i, y _i, in formula, the Gaussian parabolic line of m article of sweep trace central point on north and south and east-west direction when X, Y are imaging spectrometer acquisition high-spectrum remote sensings.

(B9) repeat (B6) ~ (B8), make i=i+1, the Gaussian parabolic line that during calculating imaging spectrometer acquisition high-spectrum remote sensing, on m article of sweep trace, all pixels are corresponding.

(B10) repeat (B1) ~ (B9), make m=m+1, the Gaussian parabolic line that during calculating imaging spectrometer acquisition high-spectrum remote sensing, all pixels of all sweep traces are corresponding.

C. image rectification

(C1) formula is utilized calculate pixel resolution GR, wherein, be the average flying height of all sweep traces during imaging spectrometer acquisition high-spectrum remote sensing, IFOV is the instantaneous field of view angle of imaging spectrometer.

(C2) formula xsize=(xmax-xmin)/GR is utilized to calculate the size xsize of correcting image in North and South direction, wherein, when xmax represents that imaging spectrometer obtains high-spectrum remote sensing, Gaussian parabolic line corresponding to all pixels is in the maximal value of North and South direction, and the Gaussian parabolic line that when xmin represents that imaging spectrometer obtains high-spectrum remote sensing, all pixels are corresponding is in the minimum value of North and South direction.

(C3) formula ysize=(ymax-ymin)/GR is utilized to calculate the size ysize of correcting image at east-west direction, wherein, when ymax represents that imaging spectrometer obtains high-spectrum remote sensing, Gaussian parabolic line corresponding to all pixels is in the maximal value of east-west direction, and the Gaussian parabolic line that when ymin represents that imaging spectrometer obtains high-spectrum remote sensing, all pixels are corresponding is in the minimum value of east-west direction.

(C4) formula xindex=(xmax-x)/GR is utilized to calculate the line number xindex of high-spectrum remote sensing first pixel correspondence in correcting image that imaging spectrometer obtains, wherein, x is the Gaussian parabolic line of the corresponding North and South direction of first pixel; Recycle the line number that all pixels of this formulae discovery are corresponding in correcting image.

(C5) formula yindex=(y-ymin)/GR is utilized to calculate the columns yindex of high-spectrum remote sensing first pixel correspondence in correcting image that imaging spectrometer obtains, wherein, y is the Gaussian parabolic line of the corresponding east-west direction of first pixel; Recycle the columns that all pixels of this formulae discovery are corresponding in correcting image.

(C6) according to line number and the columns of all pixels of high-spectrum remote sensing correspondence in correcting image of imaging spectrometer acquisition, the gray-scale value of all pixels of a kth wave band of the high-spectrum remote sensing obtained by imaging spectrometer is assigned to the corresponding pixel of correcting image, and initial value gets k=1.

(C7) inconsistent owing to correcting front and back picture size, cause in image correcting data and there is pixel disappearance, adopt nearest-neighbor method of interpolation to eliminate.

(C8) repeat (C6) ~ (C7), make k=k+1, complete all band images and correct.

With reference to Fig. 1, the present invention is by imaging spectrometer and high-precision GPS composition high-spectrum remote-sensing imaging system.Be mounted in bottom aircraft by imaging spectrometer and high-precision GPS, the camera lens of imaging spectrometer is towards ground, and imaging spectrometer sweeps acquisition high-spectrum remote sensing along with the motion of imaging platform pushes away target area; The attitude information data of imaging platform during GPS writing scan simultaneously.

With reference to Fig. 2, the process flow diagram of the high-spectrum remote sensing geometric accurate correction method based on GPS of the present invention, comprises data importing, coordinate calculates and image rectification three steps.

Step 1, data importing

With reference to Fig. 3, the data importing process flow diagram of the high-spectrum remote sensing geometric accurate correction method based on GPS of the present invention, the concrete enforcement of this step is described below:

Step (1.1) high-spectral data has three kinds of forms, is bil, bsq and bip respectively, the hyperspectral image data that in the present invention, imaging spectrometer obtains is bil form, read high-spectral data and obtain a three-dimensional matrice, front bidimensional is space dimension, and the third dimension is spectrum dimension.

Step (1.2) reads the attitude information data of the imaging platform of GPS record, comprising: sweep trace central point longitude and latitude during imaging spectrometer scanning, angle of pitch ψ (angle of aircraft longitudinal axis and surface level, unit: degree); Roll angle ω (angle of airplane pitch axis and surface level, unit: degree); Crab angle κ (angle of airplane motion direction and positive northern position, unit: degree); Flying height H (vertical range on camera lens and ground, unit: m).

When step (1.3) checks that whether the record number of the attitude information data of GPS record obtains high-spectrum remote sensing with imaging spectrometer, number of scanning lines is consistent, if inconsistent, resampling process is carried out to the attitude information data of GPS record, make both couplings.

The angle of pitch of the attitude information data that GPS records by step (1.4), roll angle and crab angle are converted to radian.

Step 2, coordinate calculates

With reference to Fig. 4, the coordinate calculation flow chart of the high-spectrum remote sensing geometric accurate correction method based on GPS of the present invention, the concrete enforcement of this step is described below:

Intermediate variable when m article of sweep trace central point longitude and latitude is changed to Gaussian parabolic line during step (2.1) calculating imaging spectrometer acquisition high-spectrum remote sensing, initial value gets m=1.The formula calculating first intermediate variable α is: α=a × B+b × sin (2B)+c × sin (4B)+d × sin (6B); Wherein, the constant coefficient calculating α comprises a, b, c and d, a=6367558.5; B=-16036.48; C=16.828; D=-0.022; B is the latitude (radian) of m article of sweep trace central point of GPS record.The formula calculating second intermediate variable β is: β=6399698.902-21562.267 × cos ²b+108.973cos ⁴b-0.612cos ⁶b.The formula calculating the 3rd intermediate variable η is: η=0.0067385254 × cos ²b.The formula calculating the 4th intermediate variable L is: L=L ₁-L _c; Wherein, L ₁for the longitude (radian) of m article of sweep trace central point of GPS record, L _cfor the median longitudinal (radian) of the projection zone that all sweep traces of GPS record are formed.

Step (2.2) utilize following formulae discovery imaging spectrometer obtain high-spectrum remote sensing time the m article of Gaussian parabolic line X of sweep trace central point in North and South direction:

$\begin{array}{c}X=\mathrm{\α}+L^2\×\mathrm{\β}\×\sin B\×\cos B/2+L^4\×\mathrm{\β}\×\sin B{\cos }^{3}B\\ \×(5-{\tan }^{2}B+{9\mathrm{\η}}^2+4{\mathrm{\η}}^4)/24+L^6\×\mathrm{\β}\×\sin B{\cos }^{5}B\\ \×(61-58{\tan }^{2}B+{\tan }^{4}B)/720\end{array};$

Step (2.3) utilize following formulae discovery imaging spectrometer obtain high-spectrum remote sensing time the m article of Gaussian parabolic line Y of sweep trace central point on east-west direction:

$\begin{array}{c}Y=L\×\mathrm{\β}\×\cos B+L^3\×\mathrm{\β}\×{\cos }^{3}B(1-{\tan }^{2}B+{\mathrm{\η}}^2)/6\\ +L^5\×\mathrm{\β}\×{\cos }^{5}B\×(9-18{\tan }^{2}B+{\tan }^{4}B)/120+500000\end{array};$

Step (2.4) utilizes formula γ=κ+pi/2 to calculate the deflection γ of m article of sweep trace when imaging spectrometer obtains high-spectrum remote sensing, wherein, the crab angle of m article of sweep trace when κ is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record.

Step (2.5) utilizes formula D=H/cos ψ to calculate when imaging spectrometer obtains high-spectrum remote sensing m article of sweep trace central point to the distance D of corresponding ground sweep trace central point, wherein, the flying height of m article of sweep trace when H is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record, the angle of pitch of m article of sweep trace when ψ is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record.

Step (2.6) utilizes formula ω _ithe roll angle ω that during=ω-(N-1) × IFOV/2+i × IFOV calculating imaging spectrometer acquisition high-spectrum remote sensing, m article of sweep trace i-th pixel is corresponding _i, wherein, i is every a line pixel number from left to right, and initial value gets i=1; N is the detection unit number of imaging spectrometer linear array, the roll angle that when ω is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record, m article of sweep trace central point is corresponding; IFOV is the instantaneous field of view angle of imaging spectrometer.

Step (2.7) utilizes formula S _i=D × tan (ω _i), Δ x _i=S _i× cos (γ) and Δ y _i=S _i× sin (γ) to calculate when imaging spectrometer obtains high-spectrum remote sensing on m article of sweep trace i-th pixel to the distance S of m article of sweep trace central point _iand the Gaussian parabolic line component Δ x on north and south and east-west direction _iwith Δ y _i.

Step (2.8) utilizes formula x _i=X+ Δ x _iand y _i=Y+ Δ y _ithe Gaussian parabolic line x that during calculating imaging spectrometer acquisition high-spectrum remote sensing, m article of sweep trace i-th pixel is corresponding _i, y _i, wherein, the Gaussian parabolic line of m article of sweep trace central point on north and south and east-west direction when X, Y are imaging spectrometer acquisition high-spectrum remote sensings.

Step (2.9) repeats step (2.6) ~ (2.8), makes i=i+1, the Gaussian parabolic line that during calculating imaging spectrometer acquisition high-spectrum remote sensing, the m article of all pixel of sweep trace is corresponding.

Step (2.10) repeats step (2.1) ~ (2.9), makes m=m+1, the Gaussian parabolic line that during calculating imaging spectrometer acquisition high-spectrum remote sensing, all pixels of all sweep traces are corresponding.

Step 3, image rectification

With reference to Fig. 5, the image rectification process flow diagram of the high-spectrum remote sensing geometric accurate correction method based on GPS of the present invention, the concrete enforcement of this step is described below:

Step (3.1) utilizes formula calculate pixel resolution GR, wherein, be the average flying height of all sweep traces during imaging spectrometer acquisition high-spectrum remote sensing, IFOV is the instantaneous field of view angle of imaging spectrometer.

Step (3.2) utilizes formula xsize=(xmax-xmin)/GR to calculate the size xsize of correcting image in North and South direction, wherein, when xmax represents that imaging spectrometer obtains high-spectrum remote sensing, Gaussian parabolic line corresponding to all pixels is in the maximal value of North and South direction, and the Gaussian parabolic line that when xmin represents that imaging spectrometer obtains high-spectrum remote sensing, all pixels are corresponding is in the minimum value of North and South direction.

Step (3.3) utilizes formula ysize=(ymax-ymin)/GR to calculate the size ysize of correcting image at east-west direction, wherein, when ymax represents that imaging spectrometer obtains high-spectrum remote sensing, Gaussian parabolic line corresponding to all pixels is in the maximal value of east-west direction, and the Gaussian parabolic line that when ymin represents that imaging spectrometer obtains high-spectrum remote sensing, all pixels are corresponding is in the minimum value of east-west direction.

Step (3.4) utilizes formula xindex=(xmax-x)/GR to calculate the line number xindex of high-spectrum remote sensing first pixel correspondence in correcting image that imaging spectrometer obtains, wherein, x is the Gaussian parabolic line of the corresponding North and South direction of first pixel; Recycle the line number that all pixels of this formulae discovery are corresponding in correcting image.

Step (3.5) utilizes formula yindex=(y-ymin)/GR to calculate the columns yindex of high-spectrum remote sensing first pixel correspondence in correcting image that imaging spectrometer obtains, wherein, y is the Gaussian parabolic line of the corresponding east-west direction of first pixel; Recycle the columns that all pixels of this formulae discovery are corresponding in correcting image.

Step (3.6) is according to the line number and the columns that calculate all pixels of high-spectrum remote sensing correspondence in correcting image that imaging spectrometer obtains, the gray-scale value of all pixels of a kth wave band of the high-spectrum remote sensing obtained by imaging spectrometer is assigned to the corresponding pixel of correcting image, and initial value gets k=1.

Step (3.7) is inconsistent owing to correcting front and back picture size, causes in image correcting data and there is pixel disappearance, adopts nearest-neighbor method of interpolation to eliminate.

Step (3.8) repeats step (3.6) ~ (3.7), makes k=k+1, completes all band images and corrects.

Claims (1)

1., based on a high-spectrum remote sensing geometric accurate correction method of GPS, it is characterized in that comprising the steps:

A. data importing;

(A1) the high-spectrum remote sensing data of the bil form that imaging spectrometer obtains are read;

(A2) read the attitude information data of the imaging platform of GPS record, comprise longitude and latitude, angle of pitch ψ, roll angle ω, crab angle κ and flying height H;

(A3) check that the number of scanning lines whether the record number of the attitude information data of GPS record obtains high-spectrum remote sensing during with imaging spectrometer is consistent, if inconsistent, resampling process carried out to the attitude information data of GPS record, mates both making;

(A4) angle of pitch ψ, roll angle ω and crab angle κ in the attitude information data recorded by GPS are converted to Rad;

The longitude and latitude that during the imaging spectrometer acquisition high-spectrum remote sensing B. recorded by GPS, sweep trace is corresponding is converted to Gaussian parabolic line;

(B1) intermediate variable when m article of sweep trace central point longitude and latitude is changed to Gaussian parabolic line during calculating imaging spectrometer acquisition high-spectrum remote sensing, initial value gets m=1; Calculate first intermediate variable α=a × B+b × sin (2B)+c × sin (4B)+d × sin (6B); Wherein, constant coefficient a=6367558.5, b=-16036.48, c=16.828, d=-0.022, B are the latitude of m article of sweep trace central point of GPS record, calculate second intermediate variable β=6399698.902-21562.267 × cos ²b+108.973cos ⁴b-0.612cos ⁶b; Calculate the 3rd intermediate variable η=0.0067385254 × cos ²b; Calculate the 4th intermediate variable L=L ₁-L _c; Wherein, L ₁for the longitude of m article of sweep trace central point of GPS record, L _cfor the median longitudinal of the projection zone that all sweep traces of GPS record are formed;

(B2) Gaussian parabolic line of m article of sweep trace central point in North and South direction during calculating imaging spectrometer acquisition high-spectrum remote sensing $\begin{array}{c}X=\mathrm{\α}+L^2\×\mathrm{\β}\×\sin B\×\cos B/2+L^4\×\mathrm{\β}\×\sin B{\cos }^{3}B\\ \×(5-{\tan }^{2}B+9{\mathrm{\η}}^2+{4\mathrm{\η}}^4)/24+L^6\×\mathrm{\β}\×\sin B{\cos }^{5}B\\ \×(61-58{\tan }^{2}B+{\tan }^{4}B)/720\end{array};$

(B3) Gaussian parabolic line of m article of sweep trace central point on east-west direction during calculating imaging spectrometer acquisition high-spectrum remote sensing $\begin{array}{c}Y=L\×\mathrm{\β}\×\cos B+L^3\×\mathrm{\β}\×{\cos }^{3}B(1-{\tan }^{2}B+{\mathrm{\η}}^2)/6\\ +L^5\×\mathrm{\β}\×{\cos }^{5}B\×(9-18{\tan }^{2}B+{\tan }^{4}B)/120+500000\end{array};$

(B4) the deflection γ=κ+pi/2 of m article of sweep trace during imaging spectrometer acquisition high-spectrum remote sensing is calculated;

(B5) m article of sweep trace central point is calculated when imaging spectrometer obtains high-spectrum remote sensing to the distance D=H/cos ψ of corresponding ground sweep trace central point, wherein, the flying height of m article of sweep trace when H is the imaging spectrometer acquisition high-spectrum remote sensing of GPS record;

(B6) the roll angle ω that during calculating imaging spectrometer acquisition high-spectrum remote sensing, m article of sweep trace i-th pixel is corresponding _i=ω-(N-1) × IFOV/2+i × IFOV, wherein, i is every a line pixel number from left to right, and initial value gets i=1; N is the detection unit number of imaging spectrometer linear array; IFOV is the instantaneous field of view angle of imaging spectrometer;

(B7) m article of sweep trace i-th pixel is calculated when imaging spectrometer obtains high-spectrum remote sensing to the distance S of m article of sweep trace central point _i=D × tan (ω _i) and Gaussian parabolic line component Δ x on north and south and east-west direction _i=S _i× cos (γ) and Δ y _i=S _i× sin (γ);

(B8) the Gaussian parabolic line x that during calculating imaging spectrometer acquisition high-spectrum remote sensing, m article of sweep trace i-th pixel is corresponding _i=X+ Δ x _i, y _i=Y+ Δ y _i, in formula, the Gaussian parabolic line of m article of sweep trace central point on north and south and east-west direction when X, Y are imaging spectrometer acquisition high-spectrum remote sensings;

(B9) repeat (B6) ~ (B8), make i add 1, the Gaussian parabolic line that during calculating imaging spectrometer acquisition high-spectrum remote sensing, on m article of sweep trace, all pixels are corresponding;

(B10) repeat (B1) ~ (B9), make m add 1, the Gaussian parabolic line that during calculating imaging spectrometer acquisition high-spectrum remote sensing, all pixels of all sweep traces are corresponding;

C. image rectification

(C1) pixel resolution is calculated wherein, it is the average flying height of all sweep traces during imaging spectrometer acquisition high-spectrum remote sensing;

(C2) size xsize=(x max-x the min)/GR of correcting image in North and South direction is calculated, wherein, when x max represents that imaging spectrometer obtains high-spectrum remote sensing, Gaussian parabolic line corresponding to all pixels is in the maximal value of North and South direction, and the Gaussian parabolic line that when x min represents that imaging spectrometer obtains high-spectrum remote sensing, all pixels are corresponding is in the minimum value of North and South direction;

(C3) size ysize=(y the max-ymin)/GR of correcting image at east-west direction is calculated, wherein, when y max represents that imaging spectrometer obtains high-spectrum remote sensing, Gaussian parabolic line corresponding to all pixels is in the maximal value of east-west direction, and the Gaussian parabolic line that when y min represents that imaging spectrometer obtains high-spectrum remote sensing, all pixels are corresponding is in the minimum value of east-west direction;

(C4) calculate line number xindex=(x the max-x)/GR of high-spectrum remote sensing first pixel correspondence in correcting image that imaging spectrometer obtains, wherein, x is the Gaussian parabolic line of the corresponding North and South direction of first pixel; Circulation performs this step and calculates all pixels line number corresponding in correcting image;

(C5) calculate columns yindex=(y-y the min)/GR of high-spectrum remote sensing first pixel correspondence in correcting image that imaging spectrometer obtains, wherein, y is the Gaussian parabolic line of the corresponding east-west direction of first pixel; Circulation performs this step and calculates all pixels columns corresponding in correcting image;

(C6) according to line number and the columns of all pixels of high-spectrum remote sensing correspondence in correcting image of imaging spectrometer acquisition, the gray-scale value of all pixels of a kth wave band of the high-spectrum remote sensing obtained by imaging spectrometer is assigned to the corresponding pixel of correcting image, and initial value gets k=1;

(C7) pixel adopting nearest-neighbor method of interpolation to eliminate owing to correcting the inconsistent image correcting data caused of front and back picture size lacks;

(C8) repeat (C6) ~ (C7), make k add 1, complete all band images and correct.