CN103076004A - Self-adaptive window variation method of fracture characteristics of multi-baseline stereoscopic matching system - Google Patents

Abstract

The invention discloses a self-adaptive window variation method of fracture characteristics of multi-baseline stereoscopic matching system. According to the invention, a regular matching window is utilized to perform object space stereoscopic matching; earth surface fracture characteristic points can be obtained through an MVLL matching method; the fracture characteristic self-adaptive window variation method is utilized to identify the earth surface fracture characteristic areas, so that the similarity degree of the matching windows is improved through changing the positions and shapes of the matching windows, so as to improve the reliability and accuracy of the matching.

Description

The fracture characteristic self-adapting window changing method of multi-base stereo matching system

Technical field

The present invention relates to a kind of fracture characteristic self-adapting window changing method of multi-base stereo matching system.

Background technology

In photogrammetric field, the essence of Image Matching is to measure or obtain same place (feature of the same name) at different images, this with computer vision in consistent for the understanding of Image Matching, i.e. correspondence.The traditional photography measurement measures same place by operator " binocular " by stereopsis, and after entering digital photogrammetry, image Auto-matching problem is considered as photogrammetric basic task.

A large amount of fruitful researchs have been made to the Image Matching technology of single stereogram in computer vision and photogrammetric field, but still have a lot of difficult problems to fail to solve, such as geometry deformation, feature similarity, block and the problem such as earth's surface fracture.When three-dimensional scenic was projected as bidimensional image, because the degree of depth and block and caused same object to have the information dropout phenomenon in the image of different visual angles, and different objects may have similar shape, size or background, caused the Image Matching difficulty.At this moment, " morbid state is separated " appears in tradition " single three-dimensional " image matching method easily, reduced widely precision and the reliability of matching result.Therefore, Image Matching is not one " algorithm ", but needs " systematization " scheme that solves above-mentioned contradiction.

Along with the extensive utilization of large (many) degrees of overlapping digital photography, three line-scan digital cameras, a large amount of image datas can be directly used in Image Matching.On this basis, the multi-base stereo matching system has obtained fast development.This system can utilize redundant image data, reduce to greatest extent the information blind area, " the morbid state solution " of single three-dimensional (overlapping twice) coupling is converted into " determining solution " of many three-dimensional (spending overlapping) coupling more, it is the focus of current Image Matching technical research and development, also be that the traditional problem constraint is broken away from digital photogrammetry, accelerate the key of full-automatic process.

In the multi-base stereo matching system, " looking the pedal line locus model " (Multiple-view VLL, MVLL) belongs to basic object space multi-base stereo matching process model more.The MVLL model is used the basic thought of VLL model, under the geometrical constraint of image nuclear line, fully utilize many superimposed image information, calculate the related data information in the match window, solve preferably similar features, blocked traditional single stereopsis coupling bottleneck problem such as feature, greatly improved reliability and the success ratio of Image Matching.Yet, studies show that of Chinese scholars, for the high-resolution remote sensing image with intensive high-lager building, the matching result of MVLL algorithm is unsatisfactory, in the problem that the corresponding imagery zone in fracture earth's surface exists mistake to mate, greatly reduce reliability and accuracy for fracture topographical features place coupling.

Summary of the invention

The fracture characteristic self-adapting window changing method that the purpose of this invention is to provide a kind of multi-base stereo matching system, can identify fracture characteristic zone, earth's surface, and position and the shape of change match window, improve the similarity degree of match window, thereby improve reliability and the accuracy of coupling.

The present invention adopts following technical proposals:

A kind of fracture characteristic self-adapting window changing method of multi-base stereo matching system may further comprise the steps

A: selected point to be matched, utilize the rule match window to carry out the object space Stereo matching, calculate match measure, obtain the optimal match point position;

B: selecting optimal match point is the current matching point, utilizes the rule match window to carry out the MVLL coupling, analyzes the match measure result, judges that whether current matching point is earth's surface fracture characteristic point, if so, enters step C; If not, determine that then current matching point is not earth's surface fracture characteristic point, and determine whether optimal match point, if optimal match point then directly calculates final coupling elevation, finish whole calculating; Then give up the current matching point if not optimal match point, finish whole calculating;

C: carry out self-adapting window and change, divide a plurality of extended window according to direction, selected one of them extended window progressively increases match window, and calculates matching threshold, then enters step D;

D: the matching threshold that analysis meter is calculated, and calculate the coupling elevation of match point according to the matching threshold that calculates.

Described steps A is further comprising the steps of

A: the forward sight on the same photography band that obtains by the line array CCD imaging system, look and three images of backsight down;

B: under the condition of determining orientation parameter, 1 P on the given ground ₀Planimetric coordinates (X ₀, Y ₀) and approximate elevation Z ₀, approximate elevation equals the average ground elevation of this area usually;

C: keep P ₀Planimetric coordinates (X ₀, Y ₀) constant, to set big or small pixel DZ as step-length, progressively change elevation and to the projection of picture side, seek the image point position on each image;

D: calculate and the correlated measure between the picture point on each image relatively, the optimal match point position is done in the corresponding picture point choosing of correlated measure maximal value, the correlated measure computing formula is as follows:

${\mathrm{\ρ}}_{I_0{I}_1}=\frac{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}(g_{i,j}-\stackrel{\‾}{g})(g_{i+r,j+c}^{\′}-{\stackrel{\‾}{g}}_{r,c}^{\′})}{\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(g_{i,j}-\stackrel{\‾}{g})}^2\·\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{(g_{i+r,j+c}^{\′}-{\stackrel{\‾}{g}}_{r,c}^{\′})}^2}$

$\stackrel{\‾}{g}=\frac{1}{m\·n}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{g}_{i,j}$ ${\stackrel{\‾}{g}}_{r,c}^{\′}=\frac{1}{m\·n}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{g}_{i+r,j+c}^{\′}$

Wherein:

Be image I ₀And I ₁Correlated measure between the upper picture point, g _{I, j}Be image I ₀The gray-scale value of upper picture point (i, j), g ' _{I+r, j+c}Be image I ₁The gray-scale value of upper picture point (i+r, j+c), With Be the gray average of Corresponding matching window, r is image I ₁With image I ₀The horizontal ordinate difference of upper same picture point, c is image I ₁With image I ₀The ordinate difference of upper same picture point, m and n are respectively the length of match window and wide.

Described step B is further comprising the steps of

E: selecting optimal match point is the current matching point, utilizes rule window to carry out the MVLL coupling, calculates related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, ρ wherein _1/2=ρ _Max/ ρ ' _Max

F: set correlation coefficient threshold ξ ₁=0.4, the ratio threshold xi ₂=1.2, judge related coefficient maximal value ρ _Max, inferior local maximum ρ _MaxWith ratio ρ _1/2Whether satisfy simultaneously following condition 1

$\left\{\begin{array}{c}{\mathrm{\&xi;}}_1<{\mathrm{\&rho;}}_{\max }<0.9\\ {\mathrm{\&xi;}}_1<{\mathrm{\&rho;}}_{\max }^{\&prime;}<0.9\end{array}\&cup;\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>0.9\\ {\mathrm{\&rho;}}_{1/2}<{\mathrm{\&xi;}}_2\end{array}\&cup;\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>{\mathrm{\&xi;}}_1\\ {\mathrm{\&rho;}}_{\max }^{\&prime;}<{\mathrm{\&xi;}}_1\\ {\mathrm{\&rho;}}_{1/2}<{\mathrm{\&xi;}}_2+{\mathrm{\&rho;}}_{\max }\end{array}\right.\right.,\right.$

If satisfy, think that then the current matching point is earth's surface fracture characteristic point, enters step C;

If do not satisfy, determine that then current matching point is not earth's surface fracture characteristic point, and judge related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Whether satisfy following condition 2,

$\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>0.9\\ {\mathrm{\&rho;}}_{\max }^{,}<0.8\\ {\mathrm{\&rho;}}_{1/2}>{\mathrm{\&xi;}}_2\end{array}\right.$

If satisfy, determine that then current matching point is optimal match point, directly calculate final coupling elevation, finish whole computation process;

If do not satisfy, then give up the current matching point, finish whole computation process.

Described step C is further comprising the steps of

G: divide a plurality of extended window according to direction, extended window can be divided into four extended window in upper and lower, left and right or upper and lower, left and right, upper left, lower-left, upper right, eight extended window in bottom right according to direction;

H: at first selected extended window, calculate the related coefficient maximal value ρ in this extended window _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, then according to the expanded in size window size of setting, and calculate the interior related coefficient maximal value ρ of this extended window _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2If, related coefficient maximal value ρ _MaxWith ratio ρ _1/2All increase, then be not more than under the prerequisite of 1.5 times of original windows at window size, continue to enlarge window size, until related coefficient maximal value ρ _MaxWith ratio ρ _1/2Till all reducing, record related coefficient maximal value this moment ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, then enter step D.

Described step D is further comprising the steps of

I: according to the related coefficient maximal value ρ that records among the step h _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Judge,

If ρ _Max＜ξ ₁, then enter step j;

If ρ _Maxξ ₁, and related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Do not satisfy condition 1 o'clock, and then entered step k;

If related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Do not meet above-mentioned two situations, then enter step l;

J: this match point is labeled as Mismatching point, then returns step h, selected next extended window is calculated;

K: this match point is labeled as correct match point, calculates the final coupling elevation of match point, finish whole computation process;

L: according to the related coefficient maximal value ρ that records among the step h _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, calculate the coupling elevation of match point under this extended window, and the coupling elevation of the match point that calculates be added on point set { Z _d, then enter step m;

M: return step h, selected next extended window until all extended window calculating are complete, enters step n;

N: analyze point set { Z _d, if max{Z _d0, max{Z then _dIt is the final coupling elevation after the fracture characteristic of match point changes through self-adapting window.

The present invention utilizes the rule match window to carry out the object space Stereo matching, and obtain earth's surface fracture characteristic point by the MVLL matching process, then utilize fracture characteristic zone, fracture characteristic self-adapting window changing method identification earth's surface, can be by changing position and the shape of match window, improve the similarity degree of match window, thereby improve reliability and the accuracy of coupling.

Description of drawings

Fig. 1 be on the same photography band that obtains by the line array CCD imaging system of the present invention forward sight, look synoptic diagram with three images of backsight down.

Embodiment

The fracture characteristic self-adapting window changing method of multi-base stereo matching system of the present invention may further comprise the steps

A: selected point to be matched, utilize the rule match window to carry out the object space Stereo matching, calculate match measure, obtain the optimal match point position;

B: selecting optimal match point is the current matching point, utilizes the rule match window to carry out the MVLL coupling, analyzes the match measure result, judges that whether current matching point is earth's surface fracture characteristic point, if so, enters step C; If not, determine that then current matching point is not earth's surface fracture characteristic point, and determine whether optimal match point, if optimal match point then directly calculates final coupling elevation, finish whole calculating; Then give up the current matching point if not optimal match point, finish whole calculating;

C: carry out self-adapting window and change, divide a plurality of extended window according to direction, selected one of them extended window progressively increases match window, and calculates matching threshold, then enters step D;

D: the matching threshold that analysis meter is calculated, and calculate the coupling elevation of match point according to the matching threshold that calculates.

Wherein, described steps A is further comprising the steps of

A: the forward sight on the same photography band that obtains by the line array CCD imaging system, look and three images of backsight down;

B: under the condition of determining orientation parameter, 1 P on the given ground ₀Planimetric coordinates (X ₀, Y ₀) and approximate elevation Z ₀, can obtain three light shafts C ₀p ₀, C ₁p ₀, C ₂p ₀(C herein ₀, C ₁, C ₂Expression and some P ₀The relevant instantaneous centre of perspectivity), the intersection point of light shafts and each self imaging be this ground point corresponding to the picture point of each image, usually be similar to the average ground elevation that elevation equals this area;

C: because approximate elevation Z ₀Out of true, establishing its error is △ Z, then P ₀Correct elevation position should be positioned at [Z ₀-△ Z, Z ₀+ △ Z] between, keep P ₀Planimetric coordinates (X ₀, Y ₀) constant, to set big or small pixel DZ as step-length, progressively change elevation and to the projection of picture side, seek the image point position on each image;

D: calculate and the correlated measure between the picture point on each image relatively, the optimal match point position is done in the corresponding picture point choosing of correlated measure maximal value, the correlated measure computing formula is as follows:

${\mathrm{\&rho;}}_{I_0{I}_1}=\frac{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}(g_{i,j}-\stackrel{\&OverBar;}{g})(g_{i+r,j+c}^{\&prime;}-{\stackrel{\&OverBar;}{g}}_{r,c}^{\&prime;})}{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{(g_{i,j}-\stackrel{\&OverBar;}{g})}^2\&CenterDot;\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{(g_{i+r,j+c}^{\&prime;}-{\stackrel{\&OverBar;}{g}}_{r,c}^{\&prime;})}^2}$

$\stackrel{\&OverBar;}{g}=\frac{1}{m\&CenterDot;n}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{g}_{i,j}$ ${\stackrel{\&OverBar;}{g}}_{r,c}^{\&prime;}=\frac{1}{m\&CenterDot;n}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{g}_{i+r,j+c}^{\&prime;}$

Wherein:

Be image I ₀And I ₁Correlated measure between the upper picture point, g _{I, j}Be image I ₀The gray-scale value of upper picture point (i, j), g ' _{I+r, j+c}Be image I ₁The gray-scale value of upper picture point (i+r, j+c), With

Be the gray average of Corresponding matching window, r is image I ₁With image I ₀The horizontal ordinate difference of upper same picture point, c is image I ₁With image I ₀The ordinate difference of upper same picture point, m and n are respectively the length of match window and wide.

Described step B is further comprising the steps of

E: selecting optimal match point is the current matching point, utilizes rule window to carry out the MVLL coupling, calculates related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, ρ wherein _1/2=ρ _Max/ ρ ' _Max

F: set correlation coefficient threshold ξ ₁=0.4, the ratio threshold xi ₂=1.2, judge related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Whether satisfy simultaneously following condition 1

$\left\{\begin{array}{c}{\mathrm{\&xi;}}_1<{\mathrm{\&rho;}}_{\max }<0.9\\ {\mathrm{\&xi;}}_1<{\mathrm{\&rho;}}_{\max }^{\&prime;}<0.9\end{array}\&cup;\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>0.9\\ {\mathrm{\&rho;}}_{1/2}<{\mathrm{\&xi;}}_2\end{array}\&cup;\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>{\mathrm{\&xi;}}_1\\ {\mathrm{\&rho;}}_{\max }^{\&prime;}<{\mathrm{\&xi;}}_1\\ {\mathrm{\&rho;}}_{1/2}<{\mathrm{\&xi;}}_2+{\mathrm{\&rho;}}_{\max }\end{array}\right.\right.,\right.$

If satisfy, think that then the current matching point is earth's surface fracture characteristic point, enters step C;

If do not satisfy, determine that then current matching point is not earth's surface fracture characteristic point, and judge related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Whether satisfy following condition 2,

$\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>0.9\\ {\mathrm{\&rho;}}_{\max }^{,}<0.8\\ {\mathrm{\&rho;}}_{1/2}>{\mathrm{\&xi;}}_2\end{array}\right.$

If satisfy, determine that then current matching point is optimal match point, directly calculate final coupling elevation, finish whole computation process;

If do not satisfy, then give up the current matching point, finish whole computation process.

Described step C is further comprising the steps of

G: divide a plurality of extended window according to direction, extended window can be divided into four extended window in upper and lower, left and right or upper and lower, left and right, upper left, lower-left, upper right, eight extended window in bottom right according to direction;

H: at first selected extended window, calculate the related coefficient maximal value ρ in this extended window _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, then according to the expanded in size window size of setting, and calculate the interior related coefficient maximal value ρ of this extended window _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2If, related coefficient maximal value ρ _MaxWith ratio ρ _1/2All increase, then be not more than under the prerequisite of 1.5 times of original windows at window size, continue to enlarge window size, until related coefficient maximal value ρ _MaxWith ratio ρ _1/2Till all reducing, record related coefficient maximal value this moment ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, then enter step D.

Described step D is further comprising the steps of

I: according to the related coefficient maximal value ρ that records among the step h _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Judge,

If ρ _Max＜ξ ₁, then enter step j;

If ρ _Maxξ ₁, and related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Do not satisfy condition 1 o'clock, and then entered step k;

If related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Do not meet above-mentioned two situations, then enter step l;

J: this match point is labeled as Mismatching point, then returns step h, selected next extended window is calculated;

K: this match point is labeled as correct match point, calculates the final coupling elevation of match point, finish whole computation process;

L: according to the related coefficient maximal value ρ that records among the step h _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, calculate the coupling elevation of match point under this extended window, and the coupling elevation of the match point that calculates be added on point set { Z _d, then enter step m;

M: return step h, selected next extended window until all extended window calculating are complete, enters step n;

N: analyze point set { Z _d, if max{Z _d0, max{Z then _dIt is the final coupling elevation after the fracture characteristic of match point changes through self-adapting window.

Claims (5)

1. the fracture characteristic self-adapting window changing method of a multi-base stereo matching system is characterized in that: may further comprise the steps

A: selected point to be matched, utilize the rule match window to carry out the object space Stereo matching, calculate match measure, obtain the optimal match point position;

B: selecting optimal match point is the current matching point, utilizes the rule match window to carry out the MVLL coupling, analyzes the match measure result, judges that whether current matching point is earth's surface fracture characteristic point, if so, enters step C; If not, determine that then current matching point is not earth's surface fracture characteristic point, and determine whether optimal match point, if optimal match point then directly calculates final coupling elevation, finish whole calculating; Then give up the current matching point if not optimal match point, finish whole calculating;

C: carry out self-adapting window and change, divide a plurality of extended window according to direction, selected one of them extended window progressively increases match window, and calculates matching threshold, then enters step D;

D: the matching threshold that analysis meter is calculated, and calculate the coupling elevation of match point according to the matching threshold that calculates.

2. the fracture characteristic self-adapting window changing method of multi-base stereo matching system according to claim 1, it is characterized in that: described steps A is further comprising the steps of

A: the forward sight on the same photography band that obtains by the line array CCD imaging system, look and three images of backsight down;

B: under the condition of determining orientation parameter, 1 P on the given ground ₀Planimetric coordinates (X ₀, Y ₀) and approximate elevation Z ₀, approximate elevation equals the average ground elevation of this area usually;

C: keep P ₀Planimetric coordinates (X ₀, Y ₀) constant, to set big or small pixel DZ as step-length, progressively change elevation and to the projection of picture side, seek the image point position on each image;

D: calculate and the correlated measure between the picture point on each image relatively, the optimal match point position is done in the corresponding picture point choosing of correlated measure maximal value, the correlated measure computing formula is as follows:

${\mathrm{\&rho;}}_{I_0{I}_1}=\frac{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}(g_{i,j}-\stackrel{\&OverBar;}{g})(g_{i+r,j+c}^{\&prime;}-{\stackrel{\&OverBar;}{g}}_{r,c}^{\&prime;})}{\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{(g_{i,j}-\stackrel{\&OverBar;}{g})}^2\&CenterDot;\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{(g_{i+r,j+c}^{\&prime;}-{\stackrel{\&OverBar;}{g}}_{r,c}^{\&prime;})}^2}$

$\stackrel{\&OverBar;}{g}=\frac{1}{m\&CenterDot;n}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{g}_{i,j}$ ${\stackrel{\&OverBar;}{g}}_{r,c}^{\&prime;}=\frac{1}{m\&CenterDot;n}\underset{i=1}{\overset{m}{\mathrm{\&Sigma;}}}\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{g}_{i+r,j+c}^{\&prime;}$

Wherein:

Be image I ₀And I ₁Correlated measure between the upper picture point, g _{I, j}Be image I ₀The gray-scale value of upper picture point (i, j), g ' _{I+r, j+c}Be image I ₁The gray-scale value of upper picture point (i+r, j+c), With Be the gray average of Corresponding matching window, r is image I ₁With image I ₀The horizontal ordinate difference of upper same picture point, c is image I ₁With image I ₀The ordinate difference of upper same picture point, m and n are respectively the length of match window and wide.

3. the fracture characteristic self-adapting window changing method of multi-base stereo matching system according to claim 1 and 2, it is characterized in that: described step B is further comprising the steps of

E: selecting optimal match point is the current matching point, utilizes rule window to carry out the MVLL coupling, calculates related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, ρ wherein _1/2=ρ _Max/ ρ ' _Max

F: set correlation coefficient threshold ξ ₁=0.4, the ratio threshold xi ₂=1.2, judge related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Whether satisfy simultaneously following condition 1

$\left\{\begin{array}{c}{\mathrm{\&xi;}}_1<{\mathrm{\&rho;}}_{\max }<0.9\\ {\mathrm{\&xi;}}_1<{\mathrm{\&rho;}}_{\max }^{\&prime;}<0.9\end{array}\&cup;\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>0.9\\ {\mathrm{\&rho;}}_{1/2}<{\mathrm{\&xi;}}_2\end{array}\&cup;\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>{\mathrm{\&xi;}}_1\\ {\mathrm{\&rho;}}_{\max }^{\&prime;}<{\mathrm{\&xi;}}_1\\ {\mathrm{\&rho;}}_{1/2}<{\mathrm{\&xi;}}_2+{\mathrm{\&rho;}}_{\max }\end{array}\right.\right.\right.,$

If satisfy, think that then the current matching point is earth's surface fracture characteristic point, enters step C;

If do not satisfy, determine that then current matching point is not earth's surface fracture characteristic point, and judge related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Whether satisfy following condition 2,

$\left\{\begin{array}{c}{\mathrm{\&rho;}}_{\max }>0.9\\ {\mathrm{\&rho;}}_{\max }^{,}<0.9\\ {\mathrm{\&rho;}}_{1/2}>{\mathrm{\&xi;}}_2\end{array}\right.$

If satisfy, determine that then current matching point is optimal match point, directly calculate final coupling elevation, finish whole computation process;

If do not satisfy, then give up the current matching point, finish whole computation process.

4. the fracture characteristic self-adapting window changing method of multi-base stereo matching system according to claim 3, it is characterized in that: described step C is further comprising the steps of

G: divide a plurality of extended window according to direction, extended window can be divided into four extended window in upper and lower, left and right or upper and lower, left and right, upper left, lower-left, upper right, eight extended window in bottom right according to direction;

H: at first selected extended window, calculate the related coefficient maximal value ρ in this extended window _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, then according to the expanded in size window size of setting, and calculate the interior related coefficient maximal value ρ of this extended window _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2If, related coefficient maximal value ρ _MaxWith ratio ρ _1/2All increase, then be not more than under the prerequisite of 1.5 times of original windows at window size, continue to enlarge window size, until related coefficient maximal value ρ _MaxWith ratio ρ _1/2Till all reducing, record related coefficient maximal value this moment ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, then enter step D.

5. the fracture characteristic self-adapting window changing method of multi-base stereo matching system according to claim 4, it is characterized in that: described step D is further comprising the steps of

I: according to the related coefficient maximal value ρ that records among the step h _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Judge,

If ρ _Max＜ξ ₁, then enter step j;

If ρ _Maxξ ₁, and related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Do not satisfy condition 1 o'clock, and then entered step k;

If related coefficient maximal value ρ _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2Do not meet above-mentioned two situations, then enter step l;

J: this match point is labeled as Mismatching point, then returns step h, selected next extended window is calculated;

K: this match point is labeled as correct match point, calculates the final coupling elevation of match point, finish whole computation process;

L: according to the related coefficient maximal value ρ that records among the step h _Max, inferior local maximum ρ ' _MaxWith ratio ρ _1/2, calculate the coupling elevation of match point under this extended window, and the coupling elevation of the match point that calculates be added on point set { Z _d, then enter step m;

M: return step h, selected next extended window until all extended window calculating are complete, enters step n;

N: analyze point set { Z _d, if max{Z _d0, max{Z then _dIt is the final coupling elevation after the fracture characteristic of match point changes through self-adapting window.