CN106920235A - Star-loaded optical remote sensing image automatic correction method based on the matching of vector base map - Google Patents

Abstract

The invention provides a kind of star-loaded optical remote sensing image automatic correction method based on the matching of vector base map, it is characterised in that including：Step A, geocoding is carried out by image to be corrected according to given resolution；Step B, polar plot is converted to the resolution ratio of binaryzation grating image, unification image to be corrected and polar plot by the given resolution；Step C, using polar plot as control base map, according to image to be corrected and control base map geography information determine image between common region；Step D, in the common region that above-mentioned steps C determines, carries out the control base map Auto-matching treatment of same resolution ratio image to be corrected and binaryzation, and the matching double points that will be obtained are used as control point calibration remote sensing image.The characteristics of present invention has good stability and positioning precision high, has expanded the usable control base map type of star-loaded optical remote sensing image automatic correction treatment, efficiently solves the problems, such as the star-loaded optical remote sensing image automatic correction without traditional optical control image overlay area.

Description

Star-loaded optical remote sensing image automatic correction method based on the matching of vector base map

Technical field

The present invention relates to remote sensing technology field, more particularly to a kind of star-loaded optical remote sensing image based on the matching of vector base map Auto-correction method.

Background technology

With succeeding in sending up for various Optical remote satellites (high score one, high score two, No. three satellites of resource etc.), remote sensing Image data amount is increasing, in order that remote sensing image data really turns into the carrier of geospatial information, supports follow-up answering With, it is necessary to assign accurate geographical location information, the accurate school of this automation to remote sensing image by each pixel of remote sensing image Positive technology proposes higher and higher requirement.At present, positioned, it is necessary to using having to realize star-loaded optical remote sensing image high-precision Control correction method realize, such method frequently with high accuracy positioning optical image as control base map, by Auto-matching Method obtains the control information of image to be corrected.But, the optics control image limited coverage area of high accuracy positioning, Er Qieyi The Position location accuracy for having data cannot guarantee that.

At present, Satellite imagery automatic geometric correction is generally completed by controlling the Auto-matching of base map with optics, solution The certainly low problem of the artificial geometric correction method low precision of remote sensing image, efficiency (Zhang Duokun, the space remote sensing figure based on images match Picture automatic, geometric and precise correction algorithm [J], remote sensing technology and application, 2006,23 (5), 545-550).

However, the automatic correction of existing Satellite imagery have the shortcomings that it is following：

(1) existing star-loaded optical remote sensing image automatic correction treatment is mostly based on optics control base map and realizes there is control The limited problem of data coverage processed.The making of optics control base map data determines that ground controls usually through manually mapping on the spot Point realize, it is necessary to expend substantial amounts of manpower and materials, the data coverage with high confidence is limited, still have a large amount of regions without Method provides optics control data, it is impossible to carry out the automatic correction process of Satellite imagery.

(2) star-loaded optical remote sensing image automatic correction processes realization using homologous raster data Auto-matching at this stage, treats Matching image is all raster data, and texture information picture is to abundant.But, vector image only records the positioning letter of interesting target Breath, different from the information that grating image is recorded, existing matching algorithm may not apply to remote sensing image and automatic of grating image With in treatment, it is necessary to design the algorithm suitable for star-loaded optical remote sensing image and vector controlled base map Auto-matching.

(3) in existing star-loaded optical remote sensing image automatic correction method, son and region often are described using feature based The matching process of statistical correlation, does not account for present in Satellite imagery geometric distortion problem in scape, in have impact on section The Position location accuracy of the heart.

In a word, the automatic correction processing method of existing Satellite imagery limited by optics control image overlay area, it is necessary to Extension can as control base map data type, and for star-loaded optical remote sensing image with vector base map the characteristics of design it is corresponding Automatic matching method, it is ensured that the stability and accuracy of automatic trimming process.

In summary, existing star-loaded optical remote sensing image automatic correction is based primarily upon optics control base map Auto-matching reality It is existing, and be based on the matching of vector base map and realize that the research that Satellite imagery is corrected automatically is few.With the increase of remotely-sensed data amount, How different control base map information is utilized, the coverage for extending automatically calibrating Satellite imagery is to need what is solved to ask Topic.

The content of the invention

(1) technical problem to be solved

In view of above-mentioned technical problem, the invention provides a kind of star-loaded optical remote sensing image based on the matching of vector base map certainly Dynamic bearing calibration, with stability is good and the characteristics of positioning precision high, has expanded star-loaded optical remote sensing image automatic correction and has processed Usable control base map type, effectively solves the star-loaded optical remote sensing image without traditional optical control image overlay area automatic Correction Problemss.

(2) technical scheme

According to an aspect of the invention, there is provided a kind of star-loaded optical remote sensing image based on the matching of vector base map is automatic Bearing calibration, including：Step A, geocoding is carried out by image to be corrected according to given resolution；Step B, institute is pressed by polar plot State the resolution ratio that given resolution is converted to binaryzation grating image, unification image to be corrected and polar plot；Step C, with vector Figure as control base map, according to image to be corrected and control base map geography information determine image between common region；Step D, In the common region that above-mentioned steps C determines, carry out same resolution ratio image to be corrected and controlled at base map Auto-matching with binaryzation Reason, the matching double points that will be obtained as control point calibration remote sensing image.

Preferably, in the step A, the RPC files that image to be corrected passes through to carry carry out geocoding, using reasonable Multinomial model determine the image coordinates (Line, Sample) of original image to be corrected and Three Dimensional Ground coordinate (Longitude, Latitude, Height) relation.

Preferably, in the step B, vector line segment in polar plot is set to 255, scalar potential by the pixel value of position Line segment is set to 0 by the pixel of position, obtain one with image to be matched with resolution ratio binaryzation grid map.

Preferably, the step D includes：Sub-step D1, selects the binaryzation grating image section of texture-rich as control System point section, and ensure that section is uniformly distributed；Sub-step D2, image to be corrected is processed using LoG detection algorithms, extracts atural object mesh Mark feature；Sub-step D3, image slice is weighted template matches after being cut into slices to vector controlled point and extracting ground object target feature； Sub-step D4, matching double points are screened using Ransac methods, remove the larger match point of error, and the matching double points for obtaining are used as control Point calibration remote sensing image processed.

Preferably, in the sub-step D1, binaryzation control base map is divided into a series of area according to the size of W*W Domain, it is that w*w and the most abundant section of texture are cut into slices as control point that a size is selected in each region.

Preferably, the abundant information degree of section is measured by counting pixel quantity of the gray value more than 0 in section.

Preferably, in the sub-step D2, image to be corrected is processed using LoG detection algorithms, extracts ground object target Center line feature as ground object target feature.

Preferably, the LoG detection algorithms include：Gaussian filtering smoothed image noise, Laplacian operators calculate second order Enhancing wire structural information is led, zero cross point is led by second order and is detected that linear structure and linear interpolation improve linear structure and position Precision.

Preferably, in the LoG detection algorithms, size and image resolution information according to actual ground object target are chosen LoG detective operators yardsticks.

Preferably, weighted template matching formula is as follows：

Wherein, C_coef(u, v) represents the likeness coefficient of template matches, and u, v represents the coordinate of sliding window, and x, y is represented Template matches centre coordinate, I and T represent reference picture and template image respectively, w (u, v) be with match center (x, y) distance into The weight coefficient of inverse relation：

Wherein, e is that weight coefficient adjusts constant, μ₁And μ₂It is the weighted intensity average of image I and T：

Wherein, N represents the pixel quantity included in template image.

(3) beneficial effect

From above-mentioned technical proposal as can be seen that star-loaded optical remote sensing image automatic school of the present invention based on the matching of vector base map Correction method at least has the advantages that one of them：

(1) star-loaded optical remote sensing, by given resolution geocoding, has been expanded as control map using vector base map Image automatic correction treatment usable control base map type, it is to avoid yardstick and rotational differential between image to be matched, effectively Solve the problems, such as the star-loaded optical remote sensing image automatic correction without traditional optical control image overlay area.

(2) by counting non-zero pixels number, the quick-pick texture-rich in segmented areas in binaryzation grating image Section improves matching process stability and obtains the distribution consistency degree at control point as section to be matched.

(3) size of LoG operators is determined with reference to atural object target size and image resolution ratio, different from common rim detection The object edge that algorithm is extracted, what the present invention was extracted is target's center's line, the information recorded in correspondence vector base map, is improved With result accuracy.

(4) the control point positioning accurate caused by being distorted in weighted template matching process amendment star-loaded optical remote sensing image scape Degree deviation, influence of the distortion to matching result in removal scape, it is ensured that the accuracy of correction result, improves image automatic correction essence Degree, stability is good.

Brief description of the drawings

By the way that shown in accompanying drawing, above and other purpose of the invention, feature and advantage will become apparent from.In whole accompanying drawings Identical reference indicates identical part.Deliberately accompanying drawing is not drawn by actual size equal proportion scaling, it is preferred that emphasis is show Go out purport of the invention.

Fig. 1 is the star-loaded optical remote sensing image automatic correction method flow based on the matching of vector base map of the embodiment of the present invention Figure.

Fig. 2 is the arrow of the star-loaded optical remote sensing image automatic correction method based on the matching of vector base map of the embodiment of the present invention Amount slice map.

Fig. 3 is the star of the star-loaded optical remote sensing image automatic correction method based on the matching of vector base map of the embodiment of the present invention Loaded optical remote sensing image and vector section Auto-matching result figure.

Fig. 4 is the star of the star-loaded optical remote sensing image automatic correction method based on the matching of vector base map of the embodiment of the present invention Loaded optical remote sensing image rectification accuracy checking point distribution map.

Fig. 5 is the star-loaded optical remote sensing image automatic correction method treatment based on the matching of vector base map of the embodiment of the present invention Positioning precision comparison check result figure before and after number remote sensing image of high score.

Specific embodiment

To make the object, technical solutions and advantages of the present invention become more apparent, below in conjunction with specific embodiment, and reference Accompanying drawing, the present invention is described in more detail.

It should be noted that in accompanying drawing or specification description, similar or identical part all uses identical figure number.It is attached The implementation for not illustrated in figure or being described, is form known to a person of ordinary skill in the art in art.In addition, though this Text can provide the demonstration of the parameter comprising particular value, it is to be understood that parameter is without being definitely equal to corresponding value, but be able to can connect The error margin received is similar to corresponding value in design constraint.The direction term mentioned in embodiment, for example " on ", D score, "front", "rear", "left", "right" etc., are only the directions of refer to the attached drawing.Therefore, the direction term for using is for illustrating not to use To limit the scope of the invention.

The present invention provides a kind of star-loaded optical remote sensing image automatic correction method based on the matching of vector base map, with stabilization The characteristics of degree is well and positioning precision is high, has expanded the usable control base map class of star-loaded optical remote sensing image automatic correction treatment Type, effectively solves the problems, such as the star-loaded optical remote sensing image automatic correction without traditional optical control image overlay area.

To make the object, technical solutions and advantages of the present invention become more apparent, below in conjunction with specific embodiment, and reference Accompanying drawing, the present invention is described in more detail.

The invention provides a kind of star-loaded optical remote sensing image automatic correction method based on the matching of vector base map.Fig. 1 is The star-loaded optical remote sensing image automatic correction method flow chart based on the matching of vector base map of the embodiment of the present invention.Refer to Fig. 1, The star-loaded optical remote sensing image automatic correction method based on the matching of vector base map of the embodiment of the present invention, including：

Step A, geocoding is carried out by image to be corrected according to given resolution；

Step B, binaryzation grating image, unification image to be corrected and arrow are converted to by polar plot by the given resolution The resolution ratio of spirogram；

Step C, using polar plot as control base map, the geography information according to image to be corrected and control base map determines image Between common region；

Step D, in the common region that above-mentioned steps C determines, carries out same resolution ratio image to be corrected and is controlled with binaryzation Base map Auto-matching treatment, the matching double points that will be obtained as control point calibration remote sensing image.

Specifically, in the step A, the RPC files that image to be corrected passes through to carry carry out geocoding, using reasonable Multinomial model determine the image coordinates (Line, Sample) of original image to be corrected and Three Dimensional Ground coordinate (Longitude, Latitude, Height) relation：

Wherein, a_n,b_n,c_n,d_n1,2,3 ... 20) for RPC parameters, (wherein, n is that, P, L, H sit for the ground of regularization Mark, X, Y are the image coordinates of regularization, and the relation between them is shown below：

Wherein, LAT_OFF, LAT_SCALE, LONG_OFF, LONG_SCALE, HEIGHT_OFF, HEIGHT_SCALE are The regularization parameter of geographical coordinates；SAMPLE_OFF, SAMPLE_SCALE, LINE_OFF, LINE_SCALE are image coordinate Regularization parameter；Longitude, Latitude and Height represent Three Dimensional Ground coordinate (longitude, latitude and elevation)；Line and Sample represents the image coordinates (row, column) of raw video.

In the step B, polar plot is converted into binaryzation grating image by given resolution, according to specified resolution Rate, can be set to 255 by vector line segment by the pixel value of position, and scalar potential line segment is set to 0 by the pixel of position, such that it is able to Obtain one it is secondary with image to be matched with the binaryzation grid map of resolution ratio.

In the step D, carry out at same resolution ratio star-loaded optical remote sensing image and binaryzation vector base map Auto-matching Reason, including：

Sub-step D1, the binaryzation grating image section for selecting texture-rich is cut into slices as control point, and ensures that section is equal Even distribution.Binaryzation vector base map is divided into a series of region according to the size of W*W first, one is selected in each region Size is w*w and the most abundant section of texture is cut into slices as control point, has so both improve the stability of subsequent match process, Ensure that the distribution consistency degree for choosing section.Because polar plot has been converted into binaryzation grating image, therefore, it can by statistics The abundant information degree of pixel quantity measurement section of the gray value more than 0 in section.

Sub-step D2, is processed using LoG (Laplacian of the Gaussian function) detection algorithm and treats school Positive image, extracts ground object target feature.Preferably, extract ground object target center line feature as ground object target feature, from And can preferably correspond to the characters of ground object of polar plot sign.

LoG detection algorithm steps include：Gaussian filtering smoothed image noise, Laplacian operators calculate second order and lead enhancing Linear structure information, led by second order zero cross point detection linear structure, linear interpolation improve linear structure positioning precision.Due to What vector base map (such as road, river polar plot etc.) was marked is the position of target's center's line, therefore, it can utilize LoG (Laplacian of the Gaussian function) detective operators process remote sensing image, extract ground object target center line special Levy.Relative to the road edge line feature of the extractions such as Sobel operators, Canny operators, the LoG operators that the present invention is used can be effective Lift the positioning precision of matching result.The selection of LoG detective operators yardsticks needs to consider the size and image of actual ground object target Resolution information.By taking road vectors figure as an example, vector mark ground object target is urban road, and road width is ten meters or so, high Divide an image resolution to be 2 meters, then road width is 5 pixels, Filtering Template size should be chosen for 5 pixels.

Sub-step D3, image slice is weighted template matches after being cut into slices to vector controlled point and extracting line feature.

Due to the influence of the factors such as hypsography, camera distortion, can exist in certain scape in star-loaded optical remote sensing image Aberration problems, and to match the coordinate of central point as control point coordinates in automatic trimming process, therefore, the area near central point Domain information should occupy bigger weight in the matching process, and matching result can be effectively improved by weighted template matching process Accuracy.

Weighted template matching formula is as follows：

Wherein, C_coef(u, v) represents the likeness coefficient of template matches, and u, v represents the coordinate of sliding window, and x, y is represented Template matches centre coordinate, I and T represent reference picture and template image respectively, w (u, v) be with match center (x, y) distance into The weight coefficient of inverse relation：

Wherein, e is that weight coefficient adjusts constant, μ₁And μ₂It is the weighted intensity average of image I and T：

Wherein, N represents the pixel quantity included in template image.

Sub-step D4, matching double points, the larger match point of removal error, the match point for obtaining are screened using Ransac methods To as control point calibration remote sensing image.

Describe star-loaded optical remote sensing image automatic school of the present invention based on the matching of vector base map in detail with reference to example Correction method.

The L1 of Beijing area grades of spaceborne optical image product of a panel height point is chosen first, is disclosed with Beijing in 2013 Road vectors figure carries out automatic correction process as control base map, the method for then being provided using the present invention.Grating image and arrow Amount base map unification be converted to 2 meters of geocoding images of resolution ratio, by the common region of image to be corrected and vector base map according to 8000m × 8000m sizes are divided into a series of equally distributed regions, in every piece of region select size be 2000m × 2000m, the most abundant section of texture.(a) and (b) figure respectively illustrate two width vector controls of the abundant degree of different texture in Fig. 2 System point section, compared with (a) figure in Fig. 2, the information content that can be used for Auto-matching treatment in Fig. 2 in (b) figure is more, using Fig. 2 In (b) figure carry out Auto-matching and can improve the stability of Auto-matching as control section.

Within 50 meters, template matches hunting zone is set to 50/2=25 pixels to the initial alignment deviation of high score one, Number image of high score is as shown in Figure 3 with the result of vector section Auto-matching.

It is 46 that Auto-matching obtains slice of data, and the automatic correction of star-loaded optical remote sensing image is carried out based on control point, It is 38s to use Intel i7-4790CPU process times.To carrying out positioning precision using the image before and after the inventive method correction Check, altogether using 10 equally distributed checkpoints, checkpoint distribution situation in image range is as shown in Figure 4.

On the basis of vector base map, number remote sensing image of high score is carried out without control correction and the present invention based on vector bottom respectively Scheme the correction process of the star-loaded optical remote sensing image automatic correction method of matching, two kinds of different disposal methods obtain the part of image Positioning precision inspection result is as shown in Figure 5.From the inspection result of positioning precision before and after correction, it is defined by vector base map, it is high Point of No. one image is 21.3 meters without control geometric positioning accuracy, and the geometry location essence of image after treatment is corrected using the present invention It is 1.3 meters to spend, and the present invention effectively improves the positioning precision of star-loaded optical remote sensing image.

One of ordinary skill in the art it should be appreciated that above implementation is intended merely to the explanation present invention, and not As limitation of the invention, as long as within the scope of the invention, change, modification to above example will all fall of the invention Within protection domain

The foregoing description of the disclosed embodiments, enables professional and technical personnel in the field to realize or uses the application. Various modifications to these embodiments will be apparent for those skilled in the art, as defined herein General Principle can in other embodiments be realized in the case where spirit herein or scope is not departed from.Therefore, the application The embodiments shown herein is not intended to be limited to, and is to fit to and principles disclosed herein and features of novelty phase one The scope most wide for causing.

It should be noted that in accompanying drawing or specification text, the implementation for not illustrating or describing is affiliated technology Form known to a person of ordinary skill in the art, is not described in detail in field.Additionally, the above-mentioned definition to each element and method is simultaneously Various concrete structures, shape or the mode mentioned in embodiment are not limited only to, those of ordinary skill in the art can carry out letter to it Singly change or replace, for example：

(1) grid map target's center linear feature extraction is carried out using LoG detective operators, it is also possible to use Sobel operators, The edge detection algorithms such as prewitt operators, Roberts operators, Canny operators, as long as using using existing edge detection operator Grating image line feature detection is carried out, is belonged within scope.

(2) present invention control base map can be road vectors figure, it would however also be possible to employ river, coastline isovector base map, Realization of the invention is not influenceed.

Particular embodiments described above, has been carried out further in detail to the purpose of the present invention, technical scheme and beneficial effect Describe in detail bright, should be understood that and the foregoing is only specific embodiment of the invention, be not intended to limit the invention, it is all Within the spirit and principles in the present invention, any modification, equivalent substitution and improvements done etc., should be included in guarantor of the invention Within the scope of shield.

Claims (10)

1. it is a kind of based on vector base map matching star-loaded optical remote sensing image automatic correction method, it is characterised in that including：

Step A, geocoding is carried out by image to be corrected according to given resolution；

Step B, binaryzation grating image, unification image to be corrected and polar plot are converted to by polar plot by the given resolution Resolution ratio；

Step C, using polar plot as control base map, between the geography information according to image to be corrected and control base map determines image Common region；

Step D, in the common region that above-mentioned steps C determines, carries out same resolution ratio image to be corrected and controls base map with binaryzation Auto-matching treatment, the matching double points that will be obtained as control point calibration remote sensing image.

2. it is as claimed in claim 1 to be based on the star-loaded optical remote sensing image automatic correction method that vector base map is matched, its feature It is that in the step A, the RPC files that image to be corrected passes through to carry carry out geocoding, using rational polynominal model Determine the image coordinates (Line, Sample) of original image to be corrected and Three Dimensional Ground coordinate (Longitude, Latitude, Height relation)：

$Y=\frac{{\mathrm{Num}}_L(P,L,H)}{{\mathrm{Den}}_L(P,L,H)},$

$X=\frac{{\mathrm{Num}}_s(P,L,H)}{{\mathrm{Den}}_s(P,L,H)},$

$\begin{array}{c}{\mathrm{Num}}_L\left(P,L,H\right)=a_1+a_{2}L+a_{3}P+a_{4}H+a_{5}LP+a_{6}LH+a_{7}PH+a_8{L}^2+a_9{P}^2+a_{10}{H}^2+a_{11}PLH+\\ a_{12}{L}^3+a_{13}{\mathrm{LP}}^2+a_{14}{\mathrm{LH}}^2+a_{15}{L}^{2}P+a_{16}{P}^3+a_{17}{\mathrm{PH}}^2+a_{18}{L}^{2}H+a_{19}{P}^{2}H+a_{20}{H}^3\end{array},$

$\begin{array}{c}{\mathrm{Num}}_s\left(P,L,H\right)=c_1+c_{2}L+c_{3}P+c_{4}H+c_{5}LP+c_{6}LH+c_{7}PH+c_8{L}^2+c_9{P}^2+c_{10}{H}^2+c_{11}PLH+\\ c_{12}{L}^3+c_{13}{\mathrm{LP}}^2+c_{14}{\mathrm{LH}}^2+c_{15}{L}^{2}P+a_{16}{P}^3+c_{17}{\mathrm{PH}}^2+c_{18}{L}^{2}H+c_{19}{P}^{2}H+c_{20}{H}^3\end{array},$

$\begin{array}{c}{\mathrm{Den}}_L\left(P,L,H\right)=b_1+b_{2}L+b_{3}P+b_{4}H+b_{5}LP+b_{6}LH+b_{7}PH+b_8{L}^2+b_9{P}^2+b_{10}{H}^2+b_{11}PLH+\\ b_{12}{L}^3+b_{13}{\mathrm{LP}}^2+b_{14}{\mathrm{LH}}^2+b_{15}{L}^{2}P+b_{16}{P}^3+b_{17}{\mathrm{PH}}^2+b_{18}{L}^{2}H+b_{19}{P}^{2}H+b_{20}{H}^3\end{array},$

$\begin{array}{c}{\mathrm{Den}}_s\left(P,L,H\right)=d_1+d_{2}L+d_{3}P+d_{4}H+d_{5}LP+d_{6}LH+d_{7}PH+d_8{L}^2+d_9{P}^2+d_{10}{H}^2+d_{11}PLH+\\ d_{12}{L}^3+d_{13}{\mathrm{LP}}^2+d_{14}{\mathrm{LH}}^2+d_{15}{L}^{2}P+d_{16}{P}^3+d_{17}{\mathrm{PH}}^2+d_{18}{L}^{2}H+d_{19}{P}^{2}H+d_{20}{H}^3\end{array},$

$P=\frac{Latitude-LAT\_OFF}{LAT\_SCALE},$

$L=\frac{Longitude-LONG\_OFF}{LONG\_SCALE},$

$H=\frac{Height-HEIGHT\_OFF}{HEIGHT\_SCALE},$

$X=\frac{Sample-SAMP\_OFF}{SAMP\_SCALE},$

$Y=\frac{Line-LINE\_OFF}{LINE\_SCALE},$

Wherein, a_n,b_n,c_n,d_nIt is RPC parameters, wherein, n is 1,2,3 ... 20, P, and L, H are the geographical coordinates of regularization, and X, Y are The image coordinates of regularization；LAT_OFF、LAT_SCALE、LONG_OFF、LONG_SCALE、HEIGHT_OFF、HEIGHT_ SCALE is the regularization parameter of geographical coordinates；SAMPLE_OFF, SAMPLE_SCALE, LINE_OFF, LINE_SCALE are image The regularization parameter of coordinate；Longitude, Latitude and Height represent Three Dimensional Ground coordinate longitude, latitude and height respectively Journey；Line and Sample represent the image coordinates row, column of raw video respectively.

3. it is as claimed in claim 1 to be based on the star-loaded optical remote sensing image automatic correction method that vector base map is matched, its feature It is in the step B, vector line segment in polar plot to be set to 255 by the pixel value of position, scalar potential line segment passes through position The pixel put is set to 0, obtain one with image to be matched with resolution ratio binaryzation grid map.

4. as any one of claim 1-3 based on vector base map matching star-loaded optical remote sensing image automatic correction side Method, it is characterised in that the step D includes：

Sub-step D1, the binaryzation grating image section for selecting texture-rich is cut into slices as control point, and ensures uniform point of section Cloth；

Sub-step D2, image to be corrected is processed using LoG detection algorithms, extracts ground object target feature；

Sub-step D3, image slice is weighted template matches after being cut into slices to vector controlled point and extracting ground object target feature；

Sub-step D4, matching double points are screened using Ransac methods, and the larger match point of removal error, the matching double points for obtaining are made It is control point calibration remote sensing image.

5. it is as claimed in claim 4 to be based on the star-loaded optical remote sensing image automatic correction method that vector base map is matched, its feature It is in the sub-step D1, binaryzation control base map to be divided into a series of region according to the size of W*W, each area It is that w*w and the most abundant section of texture are cut into slices as control point that a size is selected in domain.

6. it is as claimed in claim 5 to be based on the star-loaded optical remote sensing image automatic correction method that vector base map is matched, its feature It is that the abundant information degree of section is measured by counting pixel quantity of the gray value more than 0 in section.

7. it is as claimed in claim 4 to be based on the star-loaded optical remote sensing image automatic correction method that vector base map is matched, its feature It is in the sub-step D2, image to be corrected to be processed using LoG detection algorithms, extracts the center line feature of ground object target As the feature of ground object target.

8. it is as claimed in claim 4 to be based on the star-loaded optical remote sensing image automatic correction method that vector base map is matched, its feature It is that the LoG detection algorithms include：Gaussian filtering smoothed image noise, Laplacian operators calculate second order and lead enhancing wire Structural information, leads zero cross point and detects that linear structure and linear interpolation improve linear structure positioning precision by second order.

9. it is as claimed in claim 8 to be based on the star-loaded optical remote sensing image automatic correction method that vector base map is matched, its feature It is that in the LoG detection algorithms, size and image resolution information according to actual ground object target choose LoG detective operators Yardstick.

10. it is as claimed in claim 4 to be based on the star-loaded optical remote sensing image automatic correction method that vector base map is matched, its feature It is that weighted template matching formula is as follows：

$C_{coef}(u,v)=\frac{\mathrm{\Σ}\left(w\right(u,v\left)I\right(x-u,y-v)-{\mathrm{\μ}}_1)\left(w\right(u,v\left)T\right(x,y)-{\mathrm{\μ}}_2)}{\sqrt{\mathrm{\Σ}{\left(w\right(u,v\left)I\right(x-u,y-v)-{\mathrm{\μ}}_1)}^2\mathrm{\Σ}{\left(w\right(u,v\left)T\right(x,y)-{\mathrm{\μ}}_2)}^2}},$

Wherein, C_coef(u, v) represents the likeness coefficient of template matches, and u, v represents the coordinate of sliding window, and x, y represents template With centre coordinate, I and T represents reference picture and template image respectively, and w (u, v) is and matches center (x, y) distance and be inversely proportional pass The weight coefficient of system：

$w(u,v)=\frac{e}{\sqrt{{(x-u)}^2+{(y-v)}^2}},$

Wherein, e is that weight coefficient adjusts constant, μ₁And μ₂It is the weighted intensity average of image I and T：

$\begin{array}{c}{\mathrm{\μ}}_1=\frac{1}{N}{\Σ}_{u,v}w(u,v)I(x-u,y-v)\\ {\mathrm{\μ}}_2=\frac{1}{N}{\Σ}_{u,v}w(u,v)T(x-u,y-v)\end{array},$

Wherein, N represents the pixel quantity included in template image.