CN102508260B - Geometric imaging construction method for side-looking medium resolution ratio satellite - Google Patents

Abstract

The invention relates to a geometric imaging construction method for a side-looking medium resolution ratio satellite. The method comprises the following step: (1) importing satellite image metadata and satellite almanac data to determine a satellite orbit and an attitude fitting equation, parameter initial value of a datalink node and other parameter initial values; (2) automatically primarily acquiring an image control point initial value according to an image metadata containing projection information of a reference; (3) manually setting at least three control points, automatically matching image control points by gray matching, automatically providing a control point missing area, and modifying distribution of control points in the control point missing area; (4) modifying distribution of control points for which the step (3) is completed, transforming coordinates, and constructing a geometric imaging model of the satellite; (5) modifying, iterating and calculating model parameters according to the acquired parameter initial values of step (1) and the geometric imaging model of step (4); and (6) outputting the corrected model result and accuracy index.

Description

A kind of how much imaging construction methods of side-looking medium resolution ratio satellite

Technical field

The invention belongs to the image processing field, relate to a kind of how much imaging construction methods of side-looking medium resolution ratio satellite.

Background technology

The when and where that disaster occurs always has great randomness, progress along with modern humans society, timely natural disaster being responded fast and corresponding assistance is provided is the effective means that reduces casualty loss, thereby the monitoring of regional disaster is become particularly important.The core mode of setting up monitoring system is to set up cover reaction time remote sensing satellite system rapidly, become today of mainstream development direction at lower-cost small satellite constellation, the intermediate-resolution small satellite constellation that uses several large field angle and be observed imaging mode with side-looking becomes as the acquisition of information mode of environment mitigation monitoring system and not only guarantees response speed but also cost-effective optimal selection.And remotely-sensed data for before need to give certain Geographic Reference information, namely by specific mode it is carried out geometric manipulations, the foundation of how much imaging models of the core of geometric manipulations.

Be set up in the image that the line array sensor of the intermediate-resolution wide visual field angle side-looking observation on the moonlet collects following some principal character is arranged: sweep limit is large, is subjected to the earth curvature image serious, owing to adopt side-looking to observe this impact more serious; Field angle is large, presents in the central projection mode fully on single scan line, and differs larger with parallel projection; Satellite platform stability is general, for lower-cost moonlet, because the various factors satellite attitude jitter is more serious, can't carry out match with parallel projection.And present stage centering low resolution image the geometric manipulations means mainly rely on classical fitting of a polynomial mode, namely carry out picture point the single mode of picture point carried out registration.But adopt the medium resolution satellite of the large field angle of line array sensor, because area coverage is large, landform and earth curvature with the non-constant of single fitting of a polynomial precision, can not satisfy the drawing demand of most applications on the impact of deformation distortion.Simultaneously its also inapplicable affined transformation simulate its imaging process, the application conditions of this model is limited in long-focus, neglects the high-resolution satellite of rink corner; Because the factor of satellite platform own, the attitude jitter pattern is comparatively complicated again, and satellite to start at data precision limited, can't as geometric model imaging process be described with direct linear transformation's model (DLT) or rational polynominal model (RFM).Describe the wide visual field angle imaging and generally can only lean on traditional collinearity equation model, but describe the collinearity equation of line array sensor to the clearly definition of variation of the outer orientation parameter of sensor, and the description that parameter is changed and answer solution and related to again a series of data demands such as reference mark distributions.

Summary of the invention

The present invention solves the existing technical matters of prior art; Provide a kind of satellite metadata that takes full advantage of to consider simultaneously earth curvature, topographic relief and sensor deformation, the distortion that the composite factors such as satellite platform shake cause, improve the precision of geometric manipulations, and proposed a kind of how much imaging construction methods of side-looking medium resolution ratio satellite that respective algorithms guarantees the integrality of resolved data.

Above-mentioned technical matters of the present invention is mainly solved by following technical proposals:

A kind of how much imaging construction methods of side-looking medium resolution ratio satellite is characterized in that, may further comprise the steps:

Step 1 imports satellite image metadata and satellite almanac data, determines parameter initial value and other parameter initial values of satellite orbit and attitude fit equation and Data-Link node;

Step 2, the image metadata that has comprised projection information according to benchmark tentatively obtains the Image Control Point initial value automatically;

Step 3, at least three reference mark of manual setting are by the method Auto-matching Image Control Point of Gray-scale Matching; Then the required condition of parameter initial value of having set according to the step 1 of setting, it is regional also automatically to propose the reference mark disappearance according to satellite orbit and the distribution of the automatic detected image of attitude fitting data node location reference mark, carries out changing of reference mark distribution to lacking the zone;

Step 4, with the changing of reference mark distribution of completing steps 3, the row-coordinate conversion of going forward side by side is namely consolidated the earth's core three-dimensional cartesian coordinate system by projection coordinate with forwarding to, is planned for P under the approximate satellite body coordinate system in conjunction with the satellite orbit fitting data again _Orb=M _Orb(t) P _ECR, and how much imaging models of structure satellite;

Step 5 is carried out changing of model parameter Iterative according to how much imaging models of the satellite in the parameter initial value that has obtained in the step 1 and the step 4;

Step 6, model result and the precision index of output after correcting.

At how much imaging construction methods of above-mentioned a kind of side-looking medium resolution ratio satellite, in the described step 1, the determining of equation and Data-Link and other parameter initial values may further comprise the steps:

Step 1.1, the three respectively position vector variations of match satellite under the solid geocentric coordinate system in ground of quadratic polynomial, wherein parameter X ₀, X ₁, X ₂, Y ₀, Y ₁, Y ₂, Z ₀, Z ₁, Z ₂By the back end in the metadata according to least-squares calculation;

Step 1.2, the three respectively velocity variations of match satellite under the solid geocentric coordinate system in ground of quadratic polynomial, wherein V parameter _X0, V _X1, V _X2, V _Y0, V _Y1, V _Y2, V _Z0, V _Z1, V _Z2By the back end in the metadata according to least-squares calculation;

Step 1.3, three Data-Links, each node comprise satellite this time be engraved in attitude angle and the party position attitude angle rate of change under the body coordinate system, every Data-Link comprises 7～9 data nodes, the swing angle back end of three directions can be expressed as respectively:

To t constantly the interpolation of certain direction attitude angle according to two 3 Hermite interpolation calculation for the λ angle, at t interpolation calculation formula constantly be:

${\mathrm{\λ}}_t={\mathrm{\λ}}_i(1+2\frac{t-t_i}{t_{i+1}-t_1}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2+{\mathrm{\λ}}_{i+1}(1+2\frac{t-t_{i+1}}{t_i-t_{i+1}}){\left(\frac{t-t_i}{t_{i+1}-t_i}\right)}^2$

$+V_{\mathrm{\λi}}(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2+V_{\mathrm{\λi}+1}(t-t_{i+1}){\left(\frac{t-t_i}{t_{i+1}{-t}_i}\right)}^2$

Step 1.4 is laid the inclination angle as the initial value of constant offset with the sensor of satellite nominal, and calculates the equivalent focal length initial value of digital sensor according to the satellite flight height h of resolution r and nominal under the star of nominal:

At how much imaging construction methods of above-mentioned a kind of side-looking medium resolution ratio satellite, in the described step 2, manual setting section reference mark is namely manual chooses 3-6 reference mark, lays respectively near central and four angles of image to be corrected.

At how much imaging construction methods of above-mentioned a kind of side-looking medium resolution ratio satellite, in the described step 3, carry out reference mark the changing that distribute and namely carry out autotelicly the reference mark is carried out in disappearance zone, reference mark in the step 3 adding, concrete operation step is as follows:

Step 3.1 is set up bipartite graph B={P, C, and E}, C are the reference mark node set, and each reference mark accounts for two nodes, and P is parameter sets to be asked, and each parameter accounts for a node, if the corresponding node c in reference mark _iWith parameter p _jRelevant, c then _iWith p _jBetween a limit e is arranged _k, otherwise boundless; Wherein have reference mark number n, number of parameters m, 0≤i≤2n, 0≤j≤m; Then the point among its P only and in the corresponding time period has the limit between the point among the C in the areas imaging among the bigraph (bipartite graph) B;

Step 3.2 is according to the maximum coupling of Hungary Algorithm search bigraph (bipartite graph) B;

Step 3.3, generate the result of maximum coupling in the markers step 3.2, wherein P concentrates the parameter point that does not find coupling to need corresponding increase reference mark, mark is with the range of control of such point, preferential reference mark of interpolation in by the zone of repeating label repeatedly, the structure that refreshes bipartite graph repeats the result who generates maximum coupling in the markers step 3.2, until find P to the complete compatible coupling of C.

How much imaging construction methods at above-mentioned a kind of side-looking medium resolution ratio satellite, in the described step 4, carry out the reference mark coordinate transform, how much imaging model principles that make up satellite can be described to: point to known topocentric vector under the satellite body coordinate system through going to picture side's coordinate system after the spatial similarity conversion, its with picture side coordinate system under the collinear vector of sensing known point corresponding diagram picture point, its equation is as follows:

$\left[\begin{array}{c}\mathrm{sample}+g\left(\mathrm{sample}\right)\\ \mathrm{line}\\ -f\end{array}\right]=M_{\mathrm{Rotation}}{P}_{\mathrm{orb}}=M_{\mathrm{Rotation}}{M}_{\mathrm{orb}}\left(t\right)\·P_{\mathrm{ECR}},$

M _Rotation

Wherein:

Angle value in the formula

Obtained by the interpolation of Data-Link described in the step 1.3, g (sample) is abnormal

Varying model, this step adopt five order polynomials to be used for the distortion that the factors such as the distortion of fit line array sensor and refractive power cause: g (x)=g ₀+ g ₁X+g ₂x ²+ g ₃x ³+ g ₄x ⁴+ g ₅x ⁵

How much imaging construction methods at above-mentioned a kind of side-looking medium resolution ratio satellite, in the described step 5, carry out changing of model parameter Iterative and comprise:

Step 6.1 is set up error equation, makes up linear system and is used for iterative;

Step 6.2, regulation satellite geometry model parameter resolve the order, loop iteration is found the solution;

Step 6.3 is carried out the reference mark according to model and parameter and is just calculated.

At how much imaging construction methods of above-mentioned a kind of side-looking medium resolution ratio satellite, the concrete operation step of described step 6.1 is as follows:

Step 6.11, according to imaging model described in the step 4, the initial error equation of setting up the correction model parameter is:

The definition G (sample)=sample+g (sample), then G ' is the first order derivative of G (sample) (sample), namely G ' (x)=1+g ₁+ 2g ₂X+3g ₃x ²+ 4g ₄x ³+ 5g ₅x ⁴

The coefficient of equation is respectively:

$l_{\mathrm{sample}}=f\frac{a_1{X}_{\mathrm{orb}}+b_1{Y}_{\mathrm{orb}}+c_1{Z}_{\mathrm{orb}}}{a_3{X}_{\mathrm{orb}}+b_3{Y}_{\mathrm{orb}}+c_3{Z}_{\mathrm{orb}}}-G\left(\mathrm{sample}\right)$

$l_{\mathrm{line}}=f\frac{a_2{X}_{\mathrm{orb}}+b_2{Y}_{\mathrm{orb}}+c_2{Z}_{\mathrm{orb}}}{a_3{X}_{\mathrm{orb}}+b_3{Y}_{\mathrm{orb}}+c_3{Z}_{\mathrm{orb}}}$

$p_0=\frac{G\left(\mathrm{sample}\right)}{f{G}^{\′}\left(\mathrm{sample}\right)};$

$p_1=-\frac{1}{G^{\′}\left(\mathrm{sample}\right)};$ $p_2=-\frac{\mathrm{sample}}{G^{\′}\left(\mathrm{sample}\right)};$

$p_3=-\frac{{\mathrm{sample}}^2}{G^{\′}\left(\mathrm{sample}\right)};$ $p_4=-\frac{{\mathrm{sample}}^3}{G^{\′}\left(\mathrm{sample}\right)};$

$p_5=-\frac{{\mathrm{sample}}^4}{G^{\′}\left(\mathrm{sample}\right)};$ $p_6=-\frac{{\mathrm{sample}}^5}{G^{\′}\left(\mathrm{sample}\right)};$

$p_7=-[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\·\frac{b_2}{G^{\′}\left(\mathrm{sample}\right)};$

$p_8=-[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\·\frac{\sin {\mathrm{\κ}}_t}{G^{\′}\left(\mathrm{sample}\right)};p_9=0;$

$p_{10}=\left\{\begin{array}{c}-(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)},t_i<t\&le;t_{i+1}\\ -(1-2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)},t_{i-1}<t\&le;t_i\end{array}\right.;$

$p_{11}=\left\{\begin{array}{c}-(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{{\sin \mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)},t_i<t\&le;t_{i+1}\\ -(1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{\sin {\mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)},t_{i-1}<t\&le;t_i\end{array}\right.$

p ₁₂=0

$p_{13}=\left\{\begin{array}{c}(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)},t_i<t\&le;t_{i+1}\\ (t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)},t_{i-1}<t\&le;t_i\end{array}\right.$

$p_{14}=\left\{\begin{array}{c}(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{\sin {\mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)},t_i<t\&le;t_{i+1}\\ (t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{\sin {\mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)},t_{i-1}<t\&le;t_i\end{array}\right.$

p ₁₅=0

q ₀=q ₁=q ₂=q ₃=q ₄=q ₅=q ₆=0；

q ₇=fb ₁+fb ₃；q ₈=-fcosκ _t；q ₉=-G(sample)；

$q_{10}=\left\{\begin{array}{c}(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2(f{b}_1+f{b}_3),t_i<t\&le;t_{i+1}\\ (1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2(f{b}_1+f{b}_3),t_{i-1}<t\&le;t_i\end{array}\right.$

$q_{11}=\left\{\begin{array}{c}-(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^{2}f\cos {\mathrm{\&kappa;}}_t,t_i<t\&le;t_{i+1}\\ -(1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2{f\cos {\mathrm{\&kappa;}}_t,t}_{i-1}<t\&le;t_i\end{array}\right.$

$q_{12}=\left\{\begin{array}{c}(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^{2}G\left(\mathrm{sample}\right),t_i<t\&le;t_{i+1}\\ (1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^{2}G\left(\mathrm{sample}\right),t_{i-1}<t\&le;t_i\end{array}\right.$

$q_{13}\left\{\begin{array}{c}(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2(f{b}_1+f{b}_3),t_i<t\&le;t_{i+1}\\ (t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2(f{b}_1+f{b}_3),t_{i-1}<t{\&le;t}_i\end{array}\right.$

$q_{14}\left\{\begin{array}{c}-(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^{2}f\cos {\mathrm{\&kappa;}}_t,t_i<t\&le;t_{i+1}\\ -(t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2{f\cos {\mathrm{\&kappa;}}_t,t}_{i-1}<t{\&le;t}_i\end{array}\right.$

$q_{15}\left\{\begin{array}{c}-(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^{2}G\left(\mathrm{sample}\right),t_i<t\&le;t_{i+1}\\ -(t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^{2}G{\left(\mathrm{sample}\right),t}_{i-1}<t{\&le;t}_i\end{array}\right.$

Easy for expressing, order

$L=\left[\begin{array}{c}{l}_{\mathrm{sample}}\\ l_{\mathrm{line}}\end{array}\right]$ $V=\left[\begin{array}{c}{v}_{\mathrm{sample}}\\ v_{\mathrm{line}}\end{array}\right]$

$A=\left[\begin{array}{cccccccccccccccc}{p}_0& p_1& p_2& p_3& p_4& p_5& p_6& p_7& p_8& p_9& p_{10}& p_{11}& p_{12}& p_{13}& p_{14}& p_{15}\\ q_0& q_1& q_2& q_3& q_4& q_5& q_6& q_7& q_8& q_9& q_{10}& q_{11}& q_{12}& q_{13}& q_{14}& q_{15}\end{array}\right]$

Then error equation can be expressed as: V=A Δ-L, then directly use least square to resolve its normal equation of correction to be:

In the formula

Valuation for Δ;

Step 6.12, when finding the solution a plurality of parameter, use single general correction solution to strengthen the reliability of resolving in the parameter calculation, can't restrain or converge to improper value with what avoid that normal equation morbid state causes, namely single iteration is introduced again one deck internal layer iteration in order to the resolving Algorithm equation in the correction iteration, and the internal layer iteration is rewritten normal equation and is:

In the formula

Be the correction of the i time calculating, I is unit matrix, and this solution is that the nothing of least square is estimated partially.

At how much imaging construction methods of above-mentioned a kind of side-looking medium resolution ratio satellite, the concrete operation step of described step 6.2 is as follows:

Step 6.21 is resolved equivalent focal length f coarse value;

Step 6.22 is resolved the angular deflection constant , ω ₀, κ ₀Coarse value;

Step 6.23 is resolved distortion model coefficient g ₀, g ₁, g ₂, g ₃, g ₄, g ₅Coarse value, simultaneously changing angular deflection constant , ω ₀, κ ₀

Step 6.24, repeated execution of steps 6.21 to step 6.23 to reference mark residual error reduces speed less than 10 ^-2

Step 6.25 is revised all angle-data chain node datas

, ω _i, V _{ω i}, κ _i, V _{κ i}

Step 6.26 is refined and is calculated distortion model coefficient g ₀, g ₁, g ₂, g ₃, g ₄, g ₅No longer diminish to putting a position residual error.

At how much imaging construction methods of above-mentioned a kind of side-looking medium resolution ratio satellite, the concrete operation step of described step 6.3 is as follows:

Step 6.31, equation is resolved as shown in the formula, sample wherein in the skeleton diagram image position of using quadratic polynomial universal model calculation level position just calculating, and line, x, y are the reference mark data, are known number, s _i, l _i(i=0,1,2 ..., 4) and be this model parameter to be asked:

$\left\{\begin{array}{c}\mathrm{sample}=s_0+s_{1}x+s_{2}y+s_3{x}^2+s_4{y}^2+s_5\mathrm{xy}\\ \mathrm{line}=l_0+l_{1}x+l_{2}y+l_3{x}^2+l_4{y}^2+l_5\mathrm{xy}\end{array}\right.;$

Step 6.32 is with the corresponding time of sweep trace coordinate of general location, t=r _tLine (r _tSweep speed for sweep trace), according to the instantaneous elements of exterior orientation of Data-Link interpolation in the model parameter

ω _t, κ _t, X _St, Y _StX _St, V _Xt, V _Yt, V _Zt

Step 6.33, with initial instantaneous elements of exterior orientation substitution imaging model, according to changing of the residual values iteration scan line position of line coordinate, computing formula is as follows:

Line _i=line _I-1+ V _Line, Be respectively two coordinate components in the topocentric coordinates that changes the satellite body coordinate system over to;

Step 6.34, with the scanning instantaneous elements of exterior orientation substitution image deformation model constantly of refining, iteration is obtained sampling point range coordinate, accurately just calculates thereby finish prototype, and it is as follows that the photogrammetric distortion model correction is counted computing formula, wherein x ₀Be acquired reference mark geographic coordinate values abscissa value:

$V_{\mathrm{sample}}=-\frac{\mathrm{sampl}{e}_i+G\left(\mathrm{sampl}{e}_i\right)-x_0}{G^{\&prime;}\left(\mathrm{sampl}{e}_i\right)},\mathrm{sampl}{e}_i={\mathrm{sample}}_{i-1}+V_{\mathrm{sample}}.$

Therefore, the present invention has following advantage: take full advantage of the satellite metadata and consider simultaneously earth curvature, topographic relief and sensor deformation, the distortion that the composite factors such as satellite platform shake cause, improve the precision of geometric manipulations, and proposed the integrality of respective algorithms assurance resolved data.

Description of drawings

Fig. 1 is schematic flow sheet of the present invention.

Fig. 2 is the single imaging synoptic diagram of wide cut satellite.

Fig. 3 is imaged image reference mark initial distribution synoptic diagram.

Fig. 4 is the reference mark residual plot (wherein residual error is amplified 10 times) of indication model output of the present invention.

Fig. 5 is the reference mark residual plot (wherein residual error is amplified 10 times) of multinomial model output.

Embodiment

Below by embodiment, and by reference to the accompanying drawings, technical scheme of the present invention is described in further detail.

Embodiment:

The wide cut side-looking satellite image of this example by deciding appearance and locating information with the satellite partial orbit with and with known geographic with reference to image, lay the accurate parameters of trying to achieve the geometric model that the present invention mentions under the prerequisite of matrix the unknown at satellite sensor, provide accurately mapping relations thereby correct for geometric exact correction.Its flow process is seen Fig. 1.

This example is chosen 30m resolution (12000 * 12000) environment mitigation moonlet 1 DBMS and (is only done overshoot and correct, do not make any geometric manipulations.Its imaging mode is that typical wide cut side-looking is to sweeping imaging, as shown in Figure 2, wherein S is projection centre, and I is the sweep trace that holds shadow, and I ' is accordingly upper thread of sweep trace, O is the principal point on the single line, O ' is accordingly millet cake of principal point, T be satellite to arriving, T ' is path), and with reference to the dem data of the Landsat7 image of utm projection information and corresponding region SRTM form as a reference, model building method and the general multinomial model of using the present invention to mention compare test.

Step 1 imports satellite image metadata and satellite almanac data, determines parameter initial value and other parameter initial values of satellite orbit and attitude fit equation and Data-Link node;

(a) xml form meta data file wherein comprises initial termination time of data acquisition, calculates sweep speed according to line number;

(b) read the GPS node data of EPH form, with quadratic polynomial respectively the match satellite position change and rate variation, fit equation is expressed as:

$\left\{\begin{array}{c}{X}_t=X_0+X_{1}t+X_2{t}^2\\ Y_t=Y_0+Y_{1}t+Y_2{t}^2\\ Z_t=Z_0+Z_{1}t+Z_2{t}^2\end{array}\right.$

And:

$\left\{\begin{array}{c}{V}_{\mathrm{Xt}}=V_{X0}+V_{X1}t+V_{X2}{t}^2\\ V_{\mathrm{Yt}}=V_{Y0}+V_{Y1}t+V_{Y2}{t}^2\\ V_{\mathrm{Zt}}=V_{Z0}+V_{Z1}t+V_{Z2}{t}^2\end{array}\right.$

(c) read the attitude angle node data of ATT form, data are directly increased in the Data-Link of attitude match.

(d) calculate the equivalent focal length initial value according to nominal resolution and nominal satellite flying height.

Step 2, the image metadata that has comprised projection information according to benchmark tentatively obtains the Image Control Point initial value automatically;

Step 3, at least three reference mark of manual setting are by the method Auto-matching Image Control Point of Gray-scale Matching; Then the required condition of parameter initial value of having set according to the step 1 of setting, distribute and the automatic disappearance zone, reference mark that proposes according to satellite orbit and the automatic detected image of attitude fitting data node location reference mark, carry out changing of reference mark distribution to lacking the zone, in the present embodiment, method Auto-matching Image Control Point by Gray-scale Matching, matching image reference mark result is totally 358 points, then according to the required condition of parameter of setting, set up the bipartite graph of expression parameter and reference mark position relationship, and provide position range and the number that needs are set up the reference mark according to the solution that bipartite graph carries out maximum coupling, and a small amount of reference mark of manual increase, the reference mark distributes and shows such as Fig. 3 after revising.This reference mark can be used for general quadratic polynomial model resolves, and calculation result picture point position residual error as shown in Figure 4.

Step 4, with the changing of reference mark distribution of completing steps 3, the row-coordinate conversion of going forward side by side is namely consolidated the earth's core three-dimensional cartesian coordinate system by projection coordinate with forwarding to, is planned for P under the approximate satellite body coordinate system in conjunction with the satellite orbit fitting data again _Orb=M _Orb(t) P _ECR, P wherein _ECRBe the reference mark ECEF-Earth Centered Earth Fixed coordinate that obtains by the Gauss projection inverse operation, M _OebBe the transition matrix described in (d), P _OrbSatellite body coordinate system coordinate for the conversion acquisition.And how much imaging models of structure satellite; Should be noted that: for distortion model: g (x)=g ₀+ g ₁X+g ₂x ²+ g ₃x ³+ g ₄x ⁴+ g ₅x ⁵Wherein, this distortion model is the match of pure mathematics angle, each parameter can be interpreted as respectively: 0 time item is main relevant with sensor translation equal error, 1 item is main, and lay the factor such as inclination with sensor relevant, 2,4 items are main relevant with the factor such as lens distortion, and 3,5 times item is mainly relevant with factors such as sensor self deformation and refractive powers.

In the present embodiment,

(a) according to the radiation transformation parameter that provides in the Landsat7 data, coupling and manual corrected Image Control Point are changed in the subsidiary Geographic Reference system of reference data, namely transfer in the plane right-angle coordinate that UTM divides band;

(b) change the some position under the UTM coordinate system over to geodesic latitude and longitude coordinates system (this example adopts the geocentric coordinate system of WGS84 ellipsoid) by the inverse transformation of projective transformation, and from dem data, read the height value of this point in the bilinear interpolation mode;

(c) think that the interpolation height value is to put a position geodetic height, consolidates the earth's core three-dimensional cartesian coordinate system with geodesic latitude and longitude coordinates with changing over to;

(d) according to the picture line line number position calculation imaging time at place, reference mark, interpolation goes out the outer orientation element of satellite of instantaneous moment, position earth rotation linear velocity in conjunction with the satellite place obtains the transition matrix that the satellite body coordinate is consolidated the earth's core three-dimensional cartesian coordinate system with being tied to, further changes ground point over to the satellite body coordinate system.Transition matrix is:

Wherein:

$\stackrel{\&RightArrow;}{X}=\frac{{V_t}^{\&prime;}}{\left|\right|{V_t}^{\&prime;}\left|\right|},$ $\stackrel{\&RightArrow;}{Y}=\frac{P_t\&times;{V_t}^{\&prime;}}{\left|\right|P_t\&times;{V_t}^{\&prime;}\left|\right|},$ $\stackrel{\&RightArrow;}{Z}=\stackrel{\&RightArrow;}{X}\&times;\stackrel{\&RightArrow;}{Y},$

The parameter that indication contains time variable in the following formula as:

V _t=(V _Xt,V _Yt,V _Zt)，P _t=(X _t,Y _t,Z _t)，V _t'=(V _Xt',V _Yt',V _Zt')，

${V_{\mathrm{Xt}}}^{\&prime;}=V_{\mathrm{Xt}}-\frac{V_P{Y}_t}{\sqrt{Y_t^2+X_t^2}}\&CenterDot;\frac{Y_t}{\left|Y_t\right|},$ ${V_{\mathrm{Yt}}}^{\&prime;}=V_{\mathrm{Yt}}-\frac{V_P{Y}_t}{\sqrt{Y_t^2+X_t^2}}\&CenterDot;\frac{X_t}{\left|X_t\right|},$

$V_P=\mathrm{\&omega;}\&CenterDot;\sqrt{Y_t^2+X_t^2},$

Wherein constant ω is the angular velocity of earth rotation:

Step 5 is carried out changing of model parameter Iterative according to how much imaging models of the satellite in the parameter initial value that has obtained in the step 1 and the step 4;

Step 6, model result and the precision index of output after correcting: result of calculation output map picture point position residual distribution figure as shown in Figure 5, error quantization is compared as shown in the table in the picture point residual error that itself and general multinomial model are resolved.

Specific embodiment described herein only is to the explanation for example of the present invention's spirit.Those skilled in the art can make various modifications or replenish or adopt similar mode to substitute described specific embodiment, but can't depart from spirit of the present invention or surmount the defined scope of appended claims.

Claims (9)

1. how much imaging construction methods of a side-looking medium resolution ratio satellite is characterized in that, may further comprise the steps:

Step 1 imports satellite image metadata and satellite almanac data, determines parameter initial value and other parameter initial values of satellite orbit and attitude fit equation and Data-Link node;

Step 2, the image metadata that has comprised projection information according to benchmark tentatively obtains the Image Control Point initial value automatically;

Step 3, at least three reference mark of manual setting are by the method Auto-matching Image Control Point of Gray-scale Matching; Then the required condition of parameter initial value of having set according to the step 1 of setting, it is regional also automatically to propose the reference mark disappearance according to satellite orbit and the distribution of the automatic detected image of attitude fitting data node location reference mark, carries out changing of reference mark distribution to lacking the zone;

Step 4, with the changing of reference mark distribution of completing steps 3, the row-coordinate conversion of going forward side by side is namely consolidated the earth's core three-dimensional cartesian coordinate system by projection coordinate with forwarding to, is planned for P under the approximate satellite body coordinate system in conjunction with the satellite orbit fitting data again _Orb=M _Orb(t) P _ECR, and how much imaging models of structure satellite;

Step 5 is carried out changing of model parameter Iterative according to how much imaging models of the satellite in the parameter initial value that has obtained in the step 1 and the step 4;

Step 6, model result and the precision index of output after correcting.

2. how much imaging construction methods of a kind of side-looking medium resolution ratio satellite according to claim 1 is characterized in that, in the described step 1, the determining of equation and Data-Link and other parameter initial values may further comprise the steps:

Step 1.1, the three respectively position vector variations of match satellite under the solid geocentric coordinate system in ground of quadratic polynomial, wherein parameter X ₀, X ₁, X ₂, Y ₀, Y ₁, Y ₂, Z ₀, Z ₁, Z ₂By the back end in the metadata according to least-squares calculation;

Step 1.2, the three respectively velocity variations of match satellite under the solid geocentric coordinate system in ground of quadratic polynomial, wherein V parameter _X0, V _X1, V _X2, V _Y0, V _Y1, V _Y2, V _Z0, V _Z1, V _Z2By the back end in the metadata according to least-squares calculation;

Step 1.3, three Data-Links, each node are engraved in attitude angle and this attitude angle rate of change under the body coordinate system when comprising satellite t, every Data-Link comprises 7～9 data nodes, and the swing angle back end of three directions can be expressed as respectively:

To t constantly the interpolation of certain direction attitude angle according to two 3 Hermite interpolation calculation for the λ angle, at t interpolation calculation formula constantly be:

${\mathrm{\&lambda;}}_t={\mathrm{\&lambda;}}_i{(1+2\frac{t-t_i}{t_{i+1}-t_i})\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2+{\mathrm{\&lambda;}}_{i+1}(1+2\frac{t-t_{i+1}}{t_i-t_{i+1}}){\left(\frac{t-t_i}{t_{i+1}-t_i}\right)}^2$

$+V_{\mathrm{\&lambda;i}}(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2+V_{\mathrm{\&lambda;i}+1}(t-t_{i+1}){\left(\frac{t-t_i}{t_{i+1}-t_i}\right)}^2$

Step 1.4 is laid the inclination angle as the initial value of constant offset with the sensor of satellite nominal, and calculates the equivalent focal length initial value of digital sensor according to the satellite flight height h of resolution r and nominal under the star of nominal: $f_0=\frac{h}{r}.$

3. how much imaging construction methods of a kind of side-looking medium resolution ratio satellite according to claim 1, it is characterized in that, in the described step 2, manual setting section reference mark is namely manual chooses 3-6 reference mark, lays respectively near central and four angles of image to be corrected.

4. how much imaging construction methods of a kind of side-looking medium resolution ratio satellite according to claim 1, it is characterized in that, in the described step 3, carry out reference mark the changing that distribute and namely carry out autotelicly the reference mark is carried out in disappearance zone, reference mark in the step 3 adding, concrete operation step is as follows:

Step 3.1 is set up bipartite graph B={P, C, and E}, C are the reference mark node set, and each reference mark accounts for two nodes, and P is parameter sets to be asked, and each parameter accounts for a node, if the corresponding node c in reference mark _iWith parameter p _jRelevant, c then _iWith p _jBetween a limit e is arranged _k, otherwise boundless; Wherein have reference mark number n, number of parameters m, 0≤i≤2n, 0≤j≤m; Then the point among its P only and in the corresponding time period has the limit between the point among the C in the areas imaging among the bigraph (bipartite graph) B;

Step 3.2 is according to the maximum coupling of Hungary Algorithm search bigraph (bipartite graph) B;

Step 3.3, generate the result of maximum coupling in the markers step 3.2, wherein P concentrates the parameter point that does not find coupling to need corresponding increase reference mark, the range of control of such point of mark, preferential reference mark of interpolation in by the zone of repeating label repeatedly, the structure that refreshes bipartite graph repeats the result who generates maximum coupling in the markers step 3.2, until find P to the complete compatible coupling of C.

5. how much imaging construction methods of a kind of side-looking medium resolution ratio satellite according to claim 1, it is characterized in that, in the described step 4, carry out the reference mark coordinate transform, how much imaging model principles that make up satellite can be described to: point to known topocentric vector under the satellite body coordinate system through going to picture side's coordinate system after the spatial similarity conversion, its with picture side coordinate system under the collinear vector of sensing known point corresponding diagram picture point, its equation is as follows:

$\left[\begin{array}{c}\mathrm{sample}+g\left(\mathrm{sample}\right)\\ \mathrm{line}\\ -f\end{array}\right]=M_{\mathrm{Rotation}}{P}_{\mathrm{orb}}=M_{\mathrm{Rotation}}{M}_{\mathrm{orb}}\left(t\right)\&CenterDot;R_{\mathrm{ECR}},$

Wherein:

Angle value in the formula

Obtained by interpolation in the step 1.3, g (sample) is distortion model, and this step adopts five order polynomials to be used for the distortion that the factors such as the distortion of fit line array sensor and refractive power cause: g (x)=g ₀+ g ₁X+g ₂x ²+ g ₃x ³+ g ₄x ⁴+ g ₅x ⁵

6. how much imaging construction methods of a kind of side-looking medium resolution ratio satellite according to claim 1 is characterized in that, in the described step 5, carry out changing of model parameter Iterative and comprise:

Step 6.1 is set up error equation, makes up linear system and is used for iterative;

Step 6.2, regulation satellite geometry model parameter resolve the order, loop iteration is found the solution;

Step 6.3 is carried out the reference mark according to model and parameter and is just calculated.

7. how much imaging construction methods of a kind of side-looking medium resolution ratio satellite according to claim 6 is characterized in that the concrete operation step of described step 6.1 is as follows:

Step 6.11, according to imaging model described in the step 4, the initial error equation of setting up the correction model parameter is:

The definition G (sample)=sample+g (sample), then G ' is the first order derivative of G (sample) (sample), namely G ' (x)=1+g ₁+ 2g ₂X+3g ₃x ²+ 4g ₄x ³+ 5g ₅x ⁴

The coefficient of equation is respectively:

$l_{\mathrm{sample}}=f\frac{a_1{X}_{\mathrm{orb}}+b_1{Y}_{\mathrm{orb}}+c_1{Z}_{\mathrm{orb}}}{a_3{X}_{\mathrm{orb}}+b_3{Y}_{\mathrm{orb}}+c_3{Z}_{\mathrm{orb}}}-G\left(\mathrm{sample}\right)$

$l_{\mathrm{line}}=f\frac{a_2{X}_{\mathrm{orb}}+b_2{Y}_{\mathrm{orb}}+c_2{Z}_{\mathrm{orb}}}{a_3{X}_{\mathrm{orb}}+b_3{Y}_{\mathrm{orb}}+c_3{Z}_{\mathrm{orb}}}$

$p_0=\frac{G\left(\mathrm{sample}\right)}{f{G}^{\&prime;}\left(\mathrm{sample}\right)};$

$p_1=-\frac{1}{G^{\&prime;}\left(\mathrm{sample}\right)};$ $p_2=-\frac{\mathrm{sample}}{G^{\&prime;}\left(\mathrm{sample}\right)};$

$p_3=-\frac{{\mathrm{sample}}^2}{G^{\&prime;}\left(\mathrm{sample}\right)};$ $p_4=-\frac{{\mathrm{sample}}^3}{G^{\&prime;}\left(\mathrm{sample}\right)};$

$p_5=-\frac{{\mathrm{sample}}^4}{G^{\&prime;}\left(\mathrm{sample}\right)};$ $p_6=-\frac{{\mathrm{sample}}^5}{G^{\&prime;}\left(\mathrm{sample}\right)};$

$p_7=-[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)};$

$p_8=-[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{{\sin \mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)};p_9=0;$

$p_{10}=\left\{\begin{array}{c}-(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)},t_i<t\&le;t_{i+1}\\ -(1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)},t_{i-1}<t\&le;t_i\end{array}\right.;$

$p_{11}=\left\{\begin{array}{c}-(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{{\sin \mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)},t_i<t\&le;t_{i+1}\\ -(1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{{\sin \mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)},t_{i-1}<t\&le;t_i\end{array}\right.$

p ₁₂=0

$p_{13}=\left\{\begin{array}{c}(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)},t_i<t\&le;t_{i+1}\\ (t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{b_2}{G^{\&prime;}\left(\mathrm{sample}\right)},t_{i-1}<t\&le;t_i\end{array}\right.$

$p_{14}=\left\{\begin{array}{c}(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{{\sin \mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)},t_i<t\&le;t_{i+1}\\ (t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f+\frac{G^2\left(\mathrm{sample}\right)}{f}]\&CenterDot;\frac{{\sin \mathrm{\&kappa;}}_t}{G^{\&prime;}\left(\mathrm{sample}\right)},t_{i-1}<t\&le;t_i\end{array}\right.$

p ₁₅＝0

q ₀=q ₁=q ₂=q ₃=q ₄=q ₅=q ₆=0；

q ₇=fb ₁+fb ₃；q ₈=-fcosκ _t；q ₉=-G(sample)；

$q_{10}=\left\{\begin{array}{c}(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2({\mathrm{fb}}_1+{\mathrm{fb}}_3),t_i<t\&le;t_{i+1}\\ (1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2({\mathrm{fb}}_1+f{b}_3),t_{i-1}<t\&le;t_i\end{array}\right.$

$q_{11}=\left\{\begin{array}{c}-(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^{2}f\cos {\mathrm{\&kappa;}}_t,t_i<t\&le;t_{i+1}\\ -(1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^{2}f\cos {\mathrm{\&kappa;}}_t,t_{i-1}<t\&le;t_i\end{array}\right.$

$q_{12}=\left\{\begin{array}{c}-(1+2\frac{t-t_i}{t_{i+1}-t_i}){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2{G\left(\mathrm{sample}\right),t}_i<t\&le;t_{i+1}\\ -(1+2\frac{t-t_i}{t_{i-1}-t_i}){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^{2}G\left(\mathrm{sample}\right){,t}_{i-1}<t\&le;t_i\end{array}\right.$

$q_{13}=\left\{\begin{array}{c}(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2[f{b}_1+f{b}_3]{,t}_i<t\&le;t_{i+1}\\ (t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2[f{b}_1+f{b}_3],t_{i-1}<t\&le;t_i\end{array}\right.$

$q_{14}=\left\{\begin{array}{c}-(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^{2}f\cos {\mathrm{\&kappa;}}_t,t_i<t\&le;t_{i+1}\\ -{(t-t_i)\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^{2}f\cos {\mathrm{\&kappa;}}_t,t_{i-1}<t\&le;t_i\end{array}\right.$

$q_{15}=\left\{\begin{array}{c}-(t-t_i){\left(\frac{t-t_{i+1}}{t_i-t_{i+1}}\right)}^2{G\left(\mathrm{sample}\right),t}_i<t\&le;t_{i+1}\\ -(t-t_i){\left(\frac{t-t_{i-1}}{t_i-t_{i-1}}\right)}^2{G\left(\mathrm{sample}\right),t}_{i-1}<t\&le;t_i\end{array}\right.$

Easy for expressing, order

$L=\left[\begin{array}{c}{l}_{\mathrm{sample}}\\ l_{\mathrm{line}}\end{array}\right]$ $V=\left[\begin{array}{c}{v}_{\mathrm{sample}}\\ v_{\mathrm{line}}\end{array}\right]$

$A=\left[\begin{array}{cccccccccccccccc}{p}_0& p_1& p_2& p_3& p_4& p_5& p_6& p_7& p_8& p_9& p_{10}& p_{11}& p_{12}& p_{13}& p_{14}& p_{15}\\ q_0& q_1& q_2& q_3& q_4& q_5& q_6& q_7& q_8& q_9& q_{10}& q_{11}& q_{12}& q_{13}& q_{14}& q_{15}\end{array}\right]$

Then error equation can be expressed as: V=A Δ-L, then directly use least square to resolve its normal equation of correction to be: $\widetilde{\mathrm{\&Delta;}}={\left(A^{T}A\right)}^{-1}{A}^{T}L,$ In the formula

Valuation for Δ;

Step 6.12, when finding the solution a plurality of parameter, use single general correction solution to strengthen the reliability of resolving in the parameter calculation, can't restrain or converge to improper value with what avoid that normal equation morbid state causes, namely single iteration is introduced again one deck internal layer iteration in order to the resolving Algorithm equation in the correction iteration, and the internal layer iteration is rewritten normal equation and is:

In the formula

Be the correction of the i time calculating, I is unit matrix, and this solution is that the nothing of least square is estimated partially.

8. how much imaging construction methods of a kind of side-looking medium resolution ratio satellite according to claim 6 is characterized in that the concrete operation step of described step 6.2 is as follows:

Step 6.21 is resolved equivalent focal length f coarse value;

Step 6.22 is resolved the angular deflection constant ω ₀, κ ₀Coarse value;

Step 6.23 is resolved distortion model coefficient g ₀, g ₁, g ₂, g ₃, g ₄, g ₅Coarse value, simultaneously changing angular deflection constant

ω ₀, κ ₀

Step 6.24, repeated execution of steps 6.21 to step 6.23 to reference mark residual error reduces speed less than 10 ^-2

Step 6.25 is revised all angle-data chain node datas

ω _i, V _{ω i}, κ _i, V _{κ i}

Step 6.26 is refined and is calculated distortion model coefficient g ₀, g ₁, g ₂, g ₃, g ₄, g ₅No longer diminish to putting a position residual error.

9. how much imaging construction methods of a kind of side-looking medium resolution ratio satellite according to claim 6 is characterized in that the concrete operation step of described step 6.3 is as follows:

Step 6.31, the skeleton diagram image position of using quadratic polynomial universal model calculation level position just calculating, resolve equation as shown in the formula:

$\left\{\begin{array}{c}\mathrm{sample}=s_0+s_{1}x+s_{2}y+s_3{x}^2+s_4{y}^2+s_5\mathrm{xy}\\ \mathrm{line}=l_0+l_{1}x+l_{2}y+l_3{x}^2+l_4{y}^2+l_5\mathrm{xy}\end{array}\right.;$

Step 6.32 is with the corresponding time of sweep trace coordinate of general location, t=r _tLine (r _tSweep speed for sweep trace), according to the instantaneous elements of exterior orientation of Data-Link interpolation in the model parameter

ω _t, κ _t, X _St, Y _StX _St, V _Xt, V _Yt, V _Zt

Step 6.33, with initial instantaneous elements of exterior orientation substitution imaging model, according to changing of the residual values iteration scan line position of line coordinate, computing formula is as follows:

Line _i=line _I-1+ V _Line,

Be respectively two coordinate components in the topocentric coordinates that changes the satellite body coordinate system over to;

Step 6.34, with the scanning instantaneous elements of exterior orientation substitution image deformation model constantly of refining, iteration is obtained sampling point range coordinate, accurately just calculates thereby finish prototype, and it is as follows that the photogrammetric distortion model correction is counted computing formula:

$V_{\mathrm{sample}}=-\frac{{\mathrm{sample}}_i+G\left({\mathrm{sample}}_i\right)-x_0}{G^{\&prime;}\left({\mathrm{sample}}_i\right)},$ sample _i=sample _i-1+V _sample。

Images