CN103630120A - Mars surface linear array image epipolar ray resampling method based on strict geometric model - Google Patents

Abstract

The invention relates to a Mars surface linear array image epipolar ray resampling method based on a strict geometric model. The method comprises the following steps: acquiring parameters related to shooting positions and attitudes of linear array cameras on Mars surface and constructing the strict geometric model; calculating the epipolar ray of image point coordinates on an original image so as to acquiring a series of discrete points on the epipolar ray; setting an average elevation surface as a projection surface and subjecting a linear array stereo image to differential rectification so as to obtain a horizontal image; converting the discrete points on the epipolar ray of the original image to the horizontal image, calculating the geometrical relationship between the epipolar ray and the horizontal image and determining the direction of the epipolar ray; and carrying out epipolar ray resampling along the direction of the epipolar ray on the horizontal image so as to obtain an approximate epipolar ray image. According to the invention, the linear array image epipolar ray resampling method combines the strict geometric model and a projection locus method together for geometric analysis of the epipolar ray and carried out differential rectification to generate the horizontal image, then epipolar ray resampling is carried out on the basis of the horizontal image generated through differential rectification so as to further eliminate influence of geometric distortion, so matching precision and reliability are improved.

Description

Martian surface linear array image core line method for resampling based on tight geometric model

Technical field

The present invention relates to martian surface HRSC camera image core line method for resampling.

Background technology

Mars is the most possible planet of finding life vestige in the solar system, and scientist even dreams of that the mankind have one day can migrate Mars.By the successful execution of the orbit detectors such as Mars whole world cruiser MGS, Mars Express MEX, Mars reconnaisance orbit device MRO and courage number, Opportunity Rover, phoenix number, the Mars landing detector task such as curious number, the mankind have had brand-new understanding to Mars.Mars topographic mapping is the basis of carrying out Mars research, and obtaining high resolving power Mars terrain data is also the prerequisite of lander safe landing.Utilize photogrammetric survey method to process and remain the current main method of obtaining Mars terrain data martian surface stereopsis data.Mars Express (the MarsExpress of European Space Agency's transmitting in 2003, MEX) detector has carried the three-dimensional line-scan digital camera HRSC of high resolving power that is exclusively used in Mars topographic mapping, by HRSC image being carried out to Photogrammetric Processing, can obtain Mars whole world high resolving power terrain data.One of key problem of HRSC image photographic Measurement and Data Processing is the high precision of linear array image, quick, dense Stereo Matching, and utilizing the constraint of core line geometry is the effective ways that address this problem.

Due to the singularity of the capable central projection of line-scan digital camera, the core line geometry relation of linear array image is more complicated than frame width formula image.The people such as Kim are based on the tight geometric model of the collinearity equation linear array image core line equation of having derived, but this formula complex forms, and be non-linear form, be difficult to practical application.Satellite linear array image core line geometry relation when the people such as Habib have studied speed and attitude and are constant.The people such as Morgan utilize parallel projection model to be studied satellite image core line method for resampling.The approximate kernel line that Zhang Zuxun utilizes fitting process to study SPOT series image is arranged.Gong Danchao has studied the linear CCD image horizontal correction method based on rational function model.Hu Fen adopts linear model to simplify core curve of the same name, has proposed a kind of satellite image approximate kernel line resampling fast algorithm of practicality, and has proposed to generate based on object space projection reference surface the method for approximate kernel line image.Zhang Yongjun has proposed a kind of linear array satellite image approximate kernel line method for resampling based on rational polynominal coefficient and object space longitude and latitude.When analysis linear array image core line geometry is related to, conventionally sensor model is simplified to or is carried out certain approximate processing, as the people such as Kim are expressed as second order polynomial form by position and attitude, the people's such as Habib core line geometry analysis is based upon speed and attitude and keeps under substantially constant prerequisite, obviously, the applicability of these core line geometry models is limited.Visible, core lineation opinion the prematurity of linear array push-broom type image, also do not form unified, general core line geometry formula or core line theoretical model, so high precision, generate linear array core line image efficiently and be still worth further investigation.

Jiang Wanshou has proposed a kind of satellite stereo image pairing approximation core line generation method based on elevation datum, but the method is set up the coordinate transformation relation of core line image and raw video based on perspective transform, and the present invention is the tight geometric transformation model based on collinearity condition equation, theory is more tight, and precision is higher.

Summary of the invention

The object of this invention is to provide a kind of martian surface linear array image core line method for resampling based on tight geometric model, with high precision, the core of production linear array efficiently line image.

For achieving the above object, the martian surface linear array image core line method for resampling step based on tight geometric model of the present invention is as follows:

(1) obtain martian surface line-scan digital camera camera site, attitude correlation parameter, based on collinearity condition equation, build the strict geometric model of image;

(2), based on tight geometric model, utilize projected footprint method to calculate the core line on left image corresponding to picpointed coordinate on left image core line, the picture point on right image on original right image;

(3) set dispersed elevation face as projecting plane, utilize strict geometric model that linear array stereopsis is carried out to differential rectify and obtain horizontal image;

(4) discrete point on raw video core line is converted on horizontal image, calculates geometric relationship the definite kernel line direction of core line on horizontal image;

(5) on horizontal image, along core line direction, carry out the resampling of core line, obtain approximate kernel line image.

Further, described step (1) neutral body camera camera site, attitude correlation parameter comprise camera focus, Pixel size, pixel calibration coordinate, camera position, attitude data.

Further, the strict geometric model of described step (1) neutral body camera image is

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{c}{X}_S\\ Y_S\\ Z_S\end{array}\right]+{\mathrm{\λR}}_{J2000}^{\mathrm{MARS}}{R}_{\mathrm{star}}^{J2000}{R}_{\mathrm{body}}^{\mathrm{star}}[\left[\begin{array}{c}{d}_x\\ d_y\\ d_z\end{array}\right]+R_{\mathrm{camera}}^{\mathrm{body}}\left[\begin{array}{c}x\\ y\\ -f\end{array}\right]]---\left(1\right)$

Wherein, (x, y) is the picpointed coordinate on linear array image, [X, Y, Z] ^tfor Mars topocentric coordinates, [X _s, Y _s, Z _s] T is the Mars probes position after precise orbit determination,

for J2000 coordinate is tied to Mars body-fixed coordinate system rotation matrix,

for the detector attitude of star sensor mensuration, for the rotation matrix of satellite body to star sensor,

for the installation matrix between camera and satellite body, [dx, dy, dz] ^tfor the eccentricity component of image center under satellite body.

Further, the tight geometric model based on collinearity condition equation in described step (3) is set up the coordinate transformation relation between horizontal image and stereopsis, and recycling ground point Inverse Projection is set up the core line geometry relation of corresponding image points.

Further, the optimum scanning line search process that in described ground point Inverse Projection, ground point is corresponding is as follows: each scan line time, elements of exterior orientation, camera focus, linear array CCD that (11) obtain the tight geometry location model of linear array image to be needed are visited first location parameter; (12) calculate the projecting plane equation of each scan line of linear array image; (13) according to the projecting plane equation of camera projection centre and each scan line, calculate the distance B between each scan line projecting plane _i, D _ifor to ground point P, neighbor scanning is capable, estimate fast; (14) first judge the distance d on ground point and article one scan line projecting plane _i, i=1, if d _ibe greater than two distance B between scan line projecting plane _i, i=1, utilizes n=d _i/ n _iformula is estimated scan line increment n; (15) distance on iteration judgement ground point and i bar scan line projecting plane, until d _i<D _iand go to (16) step; (16) utilize the scan line accurate interpolation the most contiguous with P to go out optimum scanning capable

Martian surface linear array image core line method for resampling based on tight geometric model of the present invention is the linear array image core line method for resampling with the horizontal image of differential rectify generation in conjunction with the analysis of projected footprint method core line geometry, in projected footprint method, analyze on the basis of linear array image core line geometry relation, first original linear array image is carried out to the horizontal image of differential rectify generation, geometric relationship the definite approximate kernel line direction of the core line that analysis projected footprint method draws on horizontal image, then on horizontal image, carry out core line and reset produced nucleus line image, this core line that carries out on the basis of the horizontal image of differential rectify generation resamples and can further eliminate the impact of geometric distortion, improve matching precision and reliability.And for the linear array image ground point back projection problem relating in projected footprint method and geometric correction, the capable fast search algorithm of optimum scanning based on object space projecting plane geometrical constraint has been proposed, can realize high-level efficiency, high-precision optimum scanning is capable locates fast.Utilize many group HRSC image datas to test, result shows, the core line image same place vertical parallax that method of the present invention generates is better than 1 pixel, carries out the rearrangement of core line on relative level image, can effectively change geometric distortion, is conducive to Image Matching.

Accompanying drawing explanation

Fig. 1 is frame width formula image core line geometry graph of a relation;

Fig. 2 is projected footprint normal battle array image core line geometry graph of a relation;

Fig. 3 is that HRSC camera linear array is arranged schematic diagram;

Fig. 4 utilizes SPICE storehouse to carry out time and coordinate system conversion;

Fig. 5 is that linear CCD image re-projection generates relative level image;

Fig. 6 is that horizontal image coker line direction and object coordinates are X-axis angle;

Fig. 7 is the schematic diagram of HRSC linear array image core line method for resampling of the present invention;

Fig. 8 is the capable fast search process schematic diagram of optimum scanning;

Fig. 9 is projected footprint method core line analysis overall contrast figure;

Figure 10 is the local 1:1 display comparison of projected footprint method core line analysis figure;

Figure 11 is corresponding to the core line of S2 passage picture point q in embodiment;

Figure 12 is HRSC approximate kernel line image effect figure;

Figure 13 is HRSC approximate kernel line image corresponding image points vertical parallax schematic diagram.

Embodiment

1 core line ultimate principle

1.1 frame width formula image core line geometries

As shown in Figure 1, p, q are left and right image corresponding image points to frame width formula image core line geometry relation, S ₁, S ₂for taking the photograph station coordinates, S ₁with S ₂line forms photographic base B, ground point P and S ₁, S ₂form core face S ₁s ₂p, S ₁s ₂p and left and right image intersect formation core line I _pwith I _q, I _pwith I _qfor corresponding epipolar line, corresponding image points p, q must be positioned at corresponding epipolar line I _pwith I _qon.The core line image of frame width formula image can be by original dip image re-projection is realized to horizontal image (honest image), p _nwith q _ncorrespond respectively to p, the q picpointed coordinate on horizontal image, I _pnwith I _qncorresponding to I _pwith I _qprojection on horizontal image, I _pnwith I _qnon horizontal image, be parallel to each other.On picture point on horizontal image and original dip image, picture point corresponding relation is as follows

$u_n=u\left(\frac{-f}{w}\right)=-f\frac{a_{1}x+a_{2}y-a_{3}f}{c_{1}x+c_{2}y-c_{3}f}$

$v_n=u\left(\frac{-f}{w}\right)=-f\frac{b_{1}x+b_{2}y-b_{3}f}{c_{1}x+c_{2}y-c_{3}f}---\left(3\right)$

U wherein _n, v _nfor the picpointed coordinate on horizontal image, (x, y) is corresponding picpointed coordinate on original dip image, a ₁, a ₂..., c ₃included angle cosine for each axis.The core line of frame width formula image has following characteristic:

(1) core line I _pwith I _qbe straight line;

(2) the core line I on left image _pon be a little all projected to the corresponding epipolar line I of right image _qon;

(3) two core lines that corresponding image points is corresponding are corresponding one by one, and the institute on two core lines is a little also corresponding one by one.

According to above characteristic, can carry out core line resampling produced nucleus line image to original dip image, core line image column direction vertical parallax is zero, coupling can be converted to linear search by two-dimensional search accordingly.

1.2 linear array image projecting method of loci core line geometries

Due to the singularity of the capable central projection of linear array push-broom type image, can not set up like that strict core line geometry relation by frame width formula image, conventionally use projected footprint method to analyze the core line geometry characteristic of linear array image.Projected footprint method is based on collinearity condition equation, and theory is the tightest, its ultimate principle as shown in Figure 2, ground point P(X, Y, Z) on the image of left and right, image in respectively p, q, the projection centre S (X of picture point p and ground point P and this scan line _si, Y _si, Z _si) can form ，Gai projection ray of Yi Tiao projection ray and object space surface level h _iintersect and obtain a series of intersection point P _i, by P _iback projection can obtain a series of some q to right image _i, this series of some q _iform the core curve lq of picture point p, obviously, the corresponding image points q mono-that p is ordered is positioned on core curve lq.In like manner, projection ray and the object space surface level of the picture point q on right image, projection centre S ' and ground point P formation are intersected, intersection point back projection also can be obtained to a series of point to left image, form the core curve lp of q, the corresponding image points p of q also must be positioned on lp.

The concrete formula of projected footprint method core line geometry, establishes left picture picture point p (x _l, y _l) and this scan line projection centre S (X _si, Y _si, Z _si) line form the qS of projection ray, from picture point p, through the point the light of projection centre S, under object coordinates system, can be expressed as following form

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{c}{X}_{\mathrm{si}}\\ Y_{\mathrm{si}}\\ Z_{\mathrm{si}}\end{array}\right]+\mathrm{\&lambda;}\left[\begin{array}{ccc}{r}_{11}& r_{12}& r_{13}\\ r_{21}& r_{22}& r_{23}\\ r_{31}& r_{32}& r_{33}\end{array}\right]\left[\begin{array}{c}{x}_l\\ y_l\\ -f\end{array}\right]---\left(4\right)$

Wherein, r _mn(m, n=1,2,3) are the rotation matrix coefficients that this scan line is corresponding, and λ is scale factor, by above formula, can determine the upper a series of topocentric three-dimensional coordinates of the qS of projection ray.By the ground point back projection on qS light, to right image, computing formula is as follows

$x_r=-f\frac{r_{11}^{\&prime;}(X-X_{\mathrm{sj}}^{\&prime;})+r_{21}^{\&prime;}(Y-Y_{\mathrm{sj}}^{\&prime;})+r_{31}^{\&prime;}(Z-Z_{\mathrm{sj}}^{\&prime;})}{r_{13}^{\&prime;}(X-X_{\mathrm{sj}}^{\&prime;})+r_{23}^{\&prime;}(Y-Y_{\mathrm{sj}}^{\&prime;})+r_{33}^{\&prime;}(Z-Z_{\mathrm{sj}}^{\&prime;})}$

$y_r=-\mathrm{<figure-callout\; id="12"\; label="f\; r"\; filenames="HDA00003521086100031.png"\; state="\{\{state\}\}">f</figure-callout>}\frac{r_{}^{}}{}\frac{{}_{12}^{\&prime;}(X-X_{\mathrm{sj}}^{\&prime;})+r_{22}^{\&prime;}(Y-Y_{\mathrm{sj}}^{\&prime;})+r_{32}^{\&prime;}(Z-Z_{\mathrm{sj}}^{\&prime;})}{r_{13}^{\&prime;}(X-X_{\mathrm{sj}}^{\&prime;})+r_{23}^{\&prime;}(Y-Y_{\mathrm{sj}}^{\&prime;})+r_{33}^{\&prime;}(Z-Z_{\mathrm{sj}}^{\&prime;})}---\left(5\right)$

Wherein, r' _mn(m, n=1,2,3) are rotation matrix coefficient corresponding to right image j scan line, (X' _sj, Y' _sj, Z' _sj) be right image j scan line projection centre coordinate.Formula (4), (5) simultaneous are merged and can be drawn corresponding to picture point p (x on left image _l, y _l) core line equation

$x_r=-f\frac{r_{11}^{\&prime;}A+r_{21}^{\&prime;}B+r_{31}^{\&prime;}C}{r_{13}^{\&prime;}A+r_{23}^{\&prime;}B+r_{33}^{\&prime;}C}$

（6）

$y_r=-\mathrm{<figure-callout\; id="12"\; label="f\; r"\; filenames="HDA00003521086100031.png"\; state="\{\{state\}\}">f</figure-callout>}\frac{r_{}^{}}{}\frac{{}_{12}^{\&prime;}A+r_{22}^{\&prime;}B+r_{32}^{\&prime;}C}{r_{13}^{\&prime;}A+r_{23}^{\&prime;}B+r_{33}^{\&prime;}C}$

Wherein

A=X _si-X' _sj+λ(r ₁₁·x _l+r ₁₂·y _l-r ₁₃f)

B=Y _si-Y′ _sj+λ(r ₂₁·x _l+r ₂₂·y _l-r ₂₃f)

C=Z _si-Z' _sj+λ(r ₃₁·x _l+r ₃₂·y _l-r ₂₃f) （7）

During ground point elevation on the given qS of projection ray (given Z value), can calculate proportionality factors lambda, utilize formula (6) and (7) can determine that left image is as core curvilinear path lq corresponding to p point.Under actual conditions, the x coordinate of linear array image might not be zero, provides the strict difinition (x of projected footprint method core line equation herein _r, y _r).Due to the row central projection characteristic of linear array push-broom type image, rotation matrix r _mnwith r' _mnwith scan line, change, the linear array image core line equation form drawing based on projected footprint method is complicated, and is non-linear form, and formula (6) is only suitable for core line to analyze with (7), can not be directly used in core line and resample.Core line geometry relation based on forefathers to the research of linear array image core line geometry and projected footprint method, can draw the following conclusions:

(1) there is not the core line on the strict difinition of similar frame width formula image in linear array image, and the core line of linear array image neither straight line, but similar hyperbolic curve;

(2) the interior linear array image core line of subrange is close to straight line;

(3) there is not the corresponding epipolar line pair of strict difinition in linear array image, but in subrange, have core line one to one, on this basis, the coupling of linear array image can be converted to linear search by two-dimensional search.

2 martian surface linear array image core lines resample

The Mars Express HRSC image of take is below example, introduces the martian surface linear array image core line method for resampling based on tight geometric model.

The strict geometric model of 2.1 Mars Express HRSC

Mars Express is the first mars exploration task of European Space Agency, in June, 2003 transmitting, flies to Mars in Dec, 2003.The HRSC camera carrying on Mars Express is exclusively used in Mars topographic mapping, and the high resolution image obtaining covers the Mars whole world substantially.When orbit altitude is 250Km, HRSC panchromatic wave-band image resolution can reach 10m, and HRSC camera can obtain 9 wave band linear array images simultaneously, i.e. 5 panchromatic wave-band and 4 multi light spectrum handss (red, green, blue and near infrared).HRSC camera geometric parameter is in Table 1.

Table 1HRSC camera geometric parameter

Camera parameter	Index
		Wave band number	9
Panchromatic wave-band	0°、±18.9°、±12.8°
		Multi light spectrum hands	Near infrared, red, green, blue
Camera focus/mm	175
		Pixel size/um	7
Instantaneous field of view angle/arcsec	8.2
		Linear array pixel count	5184
Quantizing bit number/bit	12
		Orbit altitude/km	250
Image resolution/m	10
		Base-height ratio	0.68/0.45

As shown in Figure 3,9 line array CCD arranged parallel of HRSC camera are on focal plane, every line array CCD has 5184 pixels, coordinate system Y-axis is identical with detector heading, X-axis is along line array CCD orientation and perpendicular to Y-axis, Z axis is perpendicular to focal plane and point to photography direction, and XYZ tri-axles form right-handed system.

The initial elements of exterior orientation that obtains HRSC relates to time and coordinate system conversion, first determines the space-time datum of Mars.To be converted to the J2000 ephemeris time time, the unification of photography measurement object space coordinate system is converted to Mars body-fixed coordinate system (MarsBody-FixedCoordinateSystem).Mars reference ellipsoid is positive spheroid, and radius is 3396.0km, and the first meridian is defined in Airy-0 meteorite crater central area, and longitude and latitude definition mode is identical with the earth.Carry out in the SPICE storehouse that time and coordinate system conversion using NAIF provide, as shown in Figure 4.SPICE storehouse provides the data such as position, attitude, satellite clock information, instrument parameter and planet Coordinate system definition of satellite.

The strict geometrical model expression of Mars Express HRSC sensor based on collinearity condition equation is as follows

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{c}{X}_S\\ Y_S\\ Z_S\end{array}\right]+{\mathrm{\&lambda;R}}_{J2000}^{\mathrm{MARS}}{R}_{\mathrm{star}}^{J2000}{R}_{\mathrm{body}}^{\mathrm{star}}[\left[\begin{array}{c}{d}_x\\ d_y\\ d_z\end{array}\right]+R_{\mathrm{camera}}^{\mathrm{body}}\left[\begin{array}{c}x\\ y\\ -f\end{array}\right]]---\left(6\right)$

Wherein, (x, y) is the picpointed coordinate on HRSC image, [X, Y, Z] ^tfor Mars topocentric coordinates, [X _s, Y _s, Z _s] ^tfor the Mars probes position after precise orbit determination,

for J2000 coordinate is tied to Mars body-fixed coordinate system rotation matrix,

for the detector attitude of star sensor mensuration, for the rotation matrix of satellite body to star sensor,

for the installation matrix between camera and satellite body, [d _x, d _y, d _z] ^tfor the eccentricity component of image center under satellite body.

2.2 linear array image differential rectifies generate relative level image

The core line of frame width formula image resamples can be by by original dip image rectification, the method to horizontal image realizes.Linear array image can not frame width formula image provide clear and definite core line geometry definition like that, although and the core line geometry formula of analyzing based on projected footprint method is tight in theory, be not easy to practical operation.For linear array image core line resampling problem, can use for reference the method for the horizontal image of frame width formula image differential rectify generation, on horizontal image, carry out the resampling of core line.For this reason, based on projected footprint method, first analyze the core line geometry relation of corresponding image points on raw video herein, then, given dispersed elevation face is as projecting plane, linear array image is carried out to differential rectify and obtain horizontal image, because horizontal image and raw video can be set up strict transformational relation, therefore the core line on original linear array image can be converted to horizontal image, analyze geometric relationship the definite kernel line direction of core line on horizontal image, then on horizontal image, carry out the resampling of core line.

As shown in Figure 5, corresponding image points p, q on raw video is corresponding to corresponding image points p ', q ' on horizontal image, the core curve lp obtaining based on projected footprint method and lq form Lp ' and Lq ' on horizontal image, and owing to having eliminated the impact of droop error, Lp ' and Lq ' are approximate is straight line.Experimental result from behind can find out, core line direction is close to detector heading, and detector heading and projecting plane X-direction have certain angle, horizontal image can be resampled along core line direction after calculating this angle.In addition, on the basis of the horizontal image of differential rectify generation, carry out core line and resample and can also further eliminate the impact of geometric distortion, improve matching precision and reliability.

2.3 horizontal image coker line geometry relationship analyses

Tight geometric model based on collinearity condition equation can be set up the coordinate transformation relation between horizontal image and raw video, and recycling ground point backprojection algorithm can be set up the core line geometry relation of corresponding image points.For ease of describing, no longer list specific formula for calculation herein, and adopt Function Mapping mode to represent transfer process

Wherein formula (7) represents the mutual conversion of picture point on left image picture point and horizontal image, is just changing and is representing by raw video picpointed coordinate (x _l, y _l) picpointed coordinate (x on calculated level image _ln, y _ln), inverse transformation represents by the picpointed coordinate (x on horizontal image _ln, y _ln) calculating raw video picpointed coordinate (x _l, y _l).Formula (8) represents the mutual conversion of picture point on right image picture point image horizontal with it.Formula (9), (10) represent to utilize ground point back projection (projected footprint method) to set up the transformational relation of left and right image picpointed coordinate and horizontal image picpointed coordinate, take left image as example, can be by the picpointed coordinate (x on left image _l, y _l) calculate the picpointed coordinate (x on horizontal image _ln, y _ln), draw topocentric coordinates (X, Y, Z), carry out ground point back projection and obtain the picpointed coordinate (x on right image _r, y _r).Above-mentioned formula is on horizontal image, to carry out same place measurement and stereotactic basis, on left and right horizontal image, measures corresponding image points (x _ln, y _ln) and (x _rn, y _rn), can be converted into the picpointed coordinate on original image, and then utilize collinearity condition equation to carry out forward intersection calculating.

Utilize the point on the core curve that formula (7)～(10) can obtain projected footprint method to be converted on horizontal image, owing to having corrected the projection error that image tilts to bring through differential rectify, core curve Lp ' and Lq ' on horizontal image are approximately straight line, by calculating the slope of Lp ' and Lq ', can draw core line direction (see figure 6), along core line direction, resample, can obtain approximate kernel line image.

2.4 generate the core line method for resampling of horizontal image based on differential rectify

Core line geometry relationship analysis based on above-mentioned projected footprint method on original image and horizontal image, for Mars Express HRSC linear array image, a kind of core line method for resampling based on the horizontal image of differential rectify generation has been proposed herein, as shown in Figure 7, Mars Express HRSC image core line resampling step is as follows for algorithm idiographic flow:

(1) utilize SPICE storehouse obtain the data such as Mars Express detector position, attitude and be converted to elements of exterior orientation, based on collinearity condition equation, build the strict geometric model of Mars Express HRSC image;

(2) utilize projected footprint method to calculate the core line of picpointed coordinate on raw video, obtain a series of discrete point on core line;

(3) set dispersed elevation face as projecting plane, utilize strict geometric model that linear array stereopsis is carried out to differential rectify and obtain horizontal image;

(4) discrete point on raw video core line is converted on horizontal image, and calculates geometric relationship the definite kernel line direction of core line on horizontal image;

(5) on horizontal image, along core line direction, carry out the resampling of core line, obtain approximate kernel line image.

3 ground point back projection fast methods

Projected footprint method core line analysis and horizontal image rectification all relate to ground point back projection, and to search for the optimum scanning that this ground point is corresponding capable for given topocentric coordinates, and calculate corresponding picpointed coordinate.Traditional linear array image ground point backprojection algorithm relates to the repeatedly interative computation of collinearity equation, and efficiency is lower.It is capable that the geometrical-restriction relation that makes full use of each scan line projecting plane of linear array image is herein located optimum scanning fast, ultimate principle is as follows: as shown in Figure 8, each scan line of a HRSC linear array image and projection centre projecting plane on can construct side space, ab is that i bar scan line array CCD first, last is visited unit, S _ifor the projection centre of this scan line, image space ab and projection centre S _iand ground point AB can form the projecting plane of i bar scan line.Can calculate equally image space cd and projection centre S _i+1and the i+1 bar scan line projecting plane of ground point CD formation.Ground point P is between i bar scan line projecting plane and i+1 bar scan line projecting plane, utilize the geometric relationship between each scan line projecting plane of object space, by the distance on judgement ground point P and each scan line projecting plane, can locate fast scan line i and the i+1 the most contiguous with ground point P, it is capable that recycling ground point P and the accurate interpolation of neighbor scanning distance in the ranks go out optimum scanning.

Ground point backprojection algorithm based on object space projecting plane geometrical constraint also needs iteration to carry out, but iterative computation is simple space analysis geometric operation, compare collinearity condition equation computation amount, and during iterative computation, can use the distance between each scan line projecting plane to estimate fast, compare traditional image space collinearity equation iterative computation repeatedly, this method only needs on a small quantity space analysis geometric operation several times, so optimum scanning line search efficiency is improved significantly.

The step of the linear array image optimum scan line searching method based on object space projecting plane geometrical constraint is as follows

Step 1: each scan line time, elements of exterior orientation, camera focus, linear array CCD that obtaining the tight geometry location model of linear array image needs are visited the parameters such as first position.

Step 2: calculate the projecting plane equation of each scan line of linear array image, computing formula is

$\left|\begin{array}{ccc}x-x_1& y-y_1& z-z_1\\ x_2-x_1& y_2-y_1& z_2-z_1\\ x_3-x_1& y_3-y_1& z_3-{\mathrm{<figure-callout\; id="1"\; label="z"\; filenames="HDA00003521086100051.png"\; state="\{\{state\}\}">z</figure-callout>}}_1\end{array}\right|=0---\left(12\right)$

Above formula is converted to the general type of space plane

Ax+By+Cz+D=0 （13）

The direction number of the normal n that wherein A, B, C are space plane, and equal zero when different.1, space M (x ₀, y ₀, z ₀) to the distance of space plane, be

$d=\frac{|{\mathrm{<figure-callout\; id="0"\; label="Ax"\; filenames="HDA00003521086100012.png,HDA00003521086100021.png"\; state="\{\{state\}\}">Ax</figure-callout>}}_0+{\mathrm{By}}_0+{\mathrm{<figure-callout\; id="0"\; label="Cz"\; filenames="HDA00003521086100012.png,HDA00003521086100021.png"\; state="\{\{state\}\}">Cz</figure-callout>}}_0+D|}{\sqrt{A^2+{\mathrm{<figure-callout\; id="2"\; label="B"\; filenames="HDA00003521086100021.png"\; state="\{\{state\}\}">B</figure-callout>}}^2+C^2}}---\left(14\right)$

Main calculated amount when this range formula is the capable fast locating algorithm iteration of optimum scanning.

Step 3: calculate the distance B between each scan line projecting plane according to the projecting plane equation of camera projection centre and each scan line _i, D ₁represent the distance between article one scan line projecting plane and second scan line projecting plane, by that analogy.D _ifor to ground point P, neighbor scanning is capable, estimate fast.

Step 4: the distance d that first judges ground point P and article one scan line projecting plane _i(i=1), if d _ibe greater than two distance B between scan line projecting plane _i(i=1), utilize following formula to estimate scan line increment n, n=d _i/ n _i.

Step 5: the distance on iteration judgement P and i bar scan line projecting plane, until d _i<D _iand go to the 6th step.

Step 6: utilize the scan line accurate interpolation the most contiguous with P to go out optimum scanning capable

$L=i+\frac{D_i}{D_i+D_{i+1}}---\left(15\right)$

4 experiments and analysis

Near choosing curiosity touch-down zone your meteorite crater of lid, HRSC image is tested, and the image capturing time is on February 9th, 2008, and approximately 15 meters of image resolutions are chosen S1 and two the passage images of S2 of base-height ratio maximum in HRSC9 passage and tested.

Ground point back projection is the basis of projected footprint method core line geometry analysis herein and the horizontal image of differential rectify generation, and the direct image of the precision of ground point backprojection algorithm and efficiency is precision and the efficiency of HRSC image core line method for resampling herein.Test is 1,000,000 points of random selection on S1 and S2 passage HRSC, and given random elevation obtains topocentric coordinates, utilize backprojection algorithm by topocentric coordinates, calculate corresponding picpointed coordinate and contrast with picpointed coordinate, thereby analyze ground point backprojection algorithm precision and efficiency.Table 2 is to utilize the comparing result of tradition based on image space collinearity condition equation iterative algorithm and this paper algorithm.

Table 2 ground point backprojection algorithm precision and efficiency contrast

In analytical table, data are known, millet cake backprojection algorithm is consistent with the method precision of tradition based on image space collinearity condition equation iteration textually, all can reach the capable location of high precision optimum scanning that is better than 0.001 pixel, and algorithm computing velocity is faster herein, practical level is higher.Utilize this ground point back projection fast method can significantly promote the counting yield of projected footprint method core line analysis and differential rectify process.

Choose corresponding image points p (1836.25 on S1, S2 image, 5124.25) and q (1951.50,5430.50), utilize projected footprint method to obtain p, the q core curve on corresponding image, when projected footprint method is analyzed core line, elevation change step is taken as 1/3rd of image resolution, 5 meters.The core line calculating is superimposed upon on original HRSC image, and core curve entire and part demonstration result is shown in respectively Fig. 9 and Figure 10, and in Fig. 9, red solid line is core curve.For analyzing the geometry site of same place and core line, in Figure 11, curve is the core curve of picture point q (1951.50,5430.50) in S1 image capturing range, and round dot is picture point p.

As shown in Figure 11, the core line that projected footprint method is obtained on original HRSC image along detector heading, on the whole still close to straight line, but as seen from the figure, core line on original HRSC image can not be used straight line to be similar to, the core line geometry relation that can not utilize projected footprint method to analyze is directly carried out the resampling of core line on original HRSC.The core line that in Figure 10, original scale shows, close to straight line, shows can be similar to and regard straight line as in subrange at HRSC original image coker line.In Figure 11, on S2 image, the corresponding image points p (1836.25,5124.25) of picture point q (1951.50,5430.50) is positioned on core curve substantially, shows that the core curve drawing along projected footprint method can search corresponding image points.

With core line image generating method of the present invention, calculate HRSC image core line direction, along core line direction, resample, S1 and S2 are formed to stereo-picture, result as shown in figure 12.On stereo-picture, measure 20 pairs of corresponding image points, analyze its vertical parallax, the results are shown in Table 3 and Figure 13.

Table 3HRSC approximate kernel line image corresponding image points vertical parallax

Period	x l	y l	x r	y r	dy
						1	235.42	1337.45	247.38	1337.49	0.04
2	1082.75	1408.77	1093.42	1408.32	-0.45
						3	1472.39	1492.32	1402.55	1492.11	-0.21
4	2032.78	2308.49	2017.59	2308.9	0.41
						5	2389.45	2692.88	2389.45	2693.03	0.15
6	2548.32	3208.45	2526.19	3208.66	0.21
						7	3039.42	3348.82	3012.76	3349.31	0.49
8	3108.46	3921.32	3108.46	3921.04	-0.28
						9	3286.39	4408.62	3243.52	4408.32	-0.3
10	3498.8	4921.77	3460.28	4921.38	-0.39
						11	3591.76	5820.57	3543.36	5820.39	-0.18
12	3683.92	6020.32	3639.78	6020.49	0.17
						13	3781.64	6508.41	3757.34	6508.76	0.35
14	3904.47	7782.57	3894.57	7782.06	-0.51
						15	3988.64	8090.31	3908.35	8090.34	0.03
16	4011.56	8842.47	3922.47	8842.28	-0.19
						17	4321.66	8917.59	4201.32	8917.33	-0.26
18	4420.68	9028.32	4410.52	9027.97	-0.35
						19	4790.92	9421.49	4756.8	9421.68	0.19
20	4890.32	9632.83	4863.45	9632.68	-0.15

The stereopsis that analysis consists of core line image and same place vertical parallax, the core line image stereo measurement that method of the present invention generates is effective, and same place vertical parallax is all less than 1 pixel, has verified feasibility and the high efficiency of the inventive method.On the basis of this method core line image, can carry out the dense Stereo Matching of HRSC image, promote precision and the reliability of dense Stereo Matching.

Claims (5)

1. martian surface linear array image core line method for resampling, is characterized in that, the step of the method is as follows:

(1) obtain martian surface line-scan digital camera camera site, attitude correlation parameter, based on collinearity condition equation, build the strict geometric model of image;

(2), based on tight geometric model, utilize projected footprint method to calculate the core line on left image corresponding to picpointed coordinate on left image core line, the picture point on right image on original right image;

(3) set dispersed elevation face as projecting plane, utilize strict geometric model that linear array stereopsis is carried out to differential rectify and obtain horizontal image;

(4) discrete point on raw video core line is converted on horizontal image, calculates geometric relationship the definite kernel line direction of core line on horizontal image;

(5) on horizontal image, along core line direction, carry out the resampling of core line, obtain approximate kernel line image.

2. martian surface linear array image core line method for resampling according to claim 1, it is characterized in that, described step (1) neutral body camera camera site, attitude correlation parameter comprise camera focus, Pixel size, pixel calibration coordinate, camera position, attitude data.

3. martian surface linear array image core line method for resampling according to claim 2, is characterized in that: the strict geometric model of described step (1) neutral body camera image is

$\left[\begin{array}{c}X\\ Y\\ Z\end{array}\right]=\left[\begin{array}{c}{X}_S\\ Y_S\\ Z_S\end{array}\right]+{\mathrm{\&lambda;R}}_{J2000}^{\mathrm{MARS}}{R}_{\mathrm{star}}^{J2000}{R}_{\mathrm{body}}^{\mathrm{star}}[\left[\begin{array}{c}{d}_x\\ d_y\\ d_z\end{array}\right]+R_{\mathrm{camera}}^{\mathrm{body}}\left[\begin{array}{c}x\\ y\\ -f\end{array}\right]]---\left(1\right)$

Wherein, (x, y) is the picpointed coordinate on linear array image, [X, Y, Z] ^tfor Mars topocentric coordinates, [X _s, Y _s, Z _s] T is the Mars probes position after precise orbit determination,

for J2000 coordinate is tied to Mars body-fixed coordinate system rotation matrix,

for the detector attitude of star sensor mensuration,

for the rotation matrix of satellite body to star sensor, for the installation matrix between camera and satellite body, [dx, dy, dz] ^tfor the eccentricity component of image center under satellite body.

4. according to the martian surface linear array image core line method for resampling described in any one in claim 1-3, it is characterized in that, tight geometric model based on collinearity condition equation in described step (3) is set up the coordinate transformation relation between horizontal image and stereopsis, and recycling ground point Inverse Projection is set up the core line geometry relation of corresponding image points.

5. martian surface linear array image core line method for resampling according to claim 4, it is characterized in that, the optimum scanning line search process that in described ground point Inverse Projection, ground point is corresponding is as follows: each scan line time, elements of exterior orientation, camera focus, linear array CCD that (11) obtain the tight geometry location model of linear array image to be needed are visited first location parameter; (12) calculate the projecting plane equation of each scan line of linear array image; (13) according to the projecting plane equation of camera projection centre and each scan line, calculate the distance B between each scan line projecting plane _i, D _ifor to ground point P, neighbor scanning is capable, estimate fast; (14) first judge the distance d on ground point and article one scan line projecting plane _i, i=1, if d _ibe greater than two distance B between scan line projecting plane _i, i=1, utilizes n=d _i/ n _iformula is estimated scan line increment n; (15) distance on iteration judgement ground point and i bar scan line projecting plane, until d _i<D _iand go to (16) step; (16) utilize the scan line accurate interpolation the most contiguous with P to go out optimum scanning capable

Images