CN109727279B - Automatic registration method of vector data and remote sensing image - Google Patents

Abstract

The open remote sensing data is processed by certain decryption according to the requirements of the security policy, and the user tries to make own vectorWhen the data is superposed with the remote sensing data, the two data are not matched. Aiming at the problem, the invention discloses an automatic registration method of vector data and a remote sensing image. The method comprises the following specific steps: 1) extracting a ternary matrix A by using a scanning line algorithm based on the vector data; 2) extracting a gray matrix G based on the remote sensing image, and obtaining an edge matrix E of the gray matrix G by using a Canny operator; 3) traversing on the three-valued matrix A by using the edge matrix E to generate a normalized edge registration factor matrix F' _Edge (ii) a 4) Traversing on the three-value matrix A by using the gray matrix G to generate a normalized gray registration factor matrix F' _{Ash of} (ii) a 5) Based on the characteristics of the registration factor, carrying out automatic registration on the vector data and the remote sensing image; 6) remote sensing image data within a specified range of the vector elements are extracted.

Description

An automatic registration method of vector data and remote sensing images

technical field

The invention belongs to the field of GIS, and in particular relates to a method for automatic registration using the consistency of remote sensing image pixel features and vector data element features.

Background technique

At present, many platforms have provided open remote sensing data services, providing good data background support for many spatial browsing and retrieval applications. Especially in applications such as land use approval, environmental assessment, and road planning, it is often necessary to extract remote sensing images within a certain range of vector elements for analysis. However, since the relevant open data has been decrypted according to the requirements of the security policy, when users try to overlay their own vector data and remote sensing data, there will be a mismatch between the two. Therefore, there is an urgent need for an efficient and automatic method to correct the geometric deviation between the self-owned vector data and the public remote sensing image data.

Due to the different properties of vector data and remote sensing images, there are certain rotation changes, scale changes and position changes. At present, the mainstream practice is to manually select certain ground control points to obtain the transformation function between them for calibration. This method has high requirements for operation accuracy, is time-consuming and labor-intensive, and has problems such as inability to download map data, inaccurate calibration of the naked eye, and inconsistent zooming degrees. It is not suitable as a conventional method for processing massive geographic information data. It cannot be used as a rigorous scientific method for high-precision, large-scale data processing and updating.

In recent years, Chinese and foreign scholars have proposed many methods for the automatic registration of remote sensing images and vector data. Joachim Hohle (2008) used the old orthophoto and the existing vector map as a reference, obtained the image template of the road intersection through the orthophoto and the vector map, matched the new image to generate new control points, and realized the automatic remote sensing image. outward orientation. Experiments show that it can basically meet the requirements of orthophoto and vector map update; Heiner Hild et al. (2001) took the registration of SPOT image and vector map as the research object, and used the initial control points selected manually to determine the approximate transformation relationship between the two. , extracting control points on the polygon boundary for aerial triangulation to realize the registration of images and maps; Zhang Xiaodong et al. (2006) performed automatic registration of remote sensing images and vector data based on polygonal features of surface objects; Liu Zhiqing (2012) used Canny Operators and edge tracking are used to filter and extract suitable linear features of remote sensing images and perform similarity measurement to realize automatic registration of remote sensing images and vector data.

The processing steps of the above method generally include: feature selection and extraction, vector data preprocessing, feature matching, and realization of vector-grid registration. These research methods are only for a specific type of problem, and most methods still require a lot of manual interaction, and there is no fully automated mature system, which is difficult to meet the needs of efficient large-scale image data processing.

SUMMARY OF THE INVENTION

The invention mainly aims at the characteristics of weak shape deformation and angular deformation between the self-owned vector data and the public remote sensing image data, and the position offset is relatively large. Automatic cropping and extraction of remote sensing image data within a certain feature or feature set. On the basis of extracting the registration factor between the two, by calculating the relative deviation between the public remote sensing image data and the user's own vector data, the automatic extraction of remote sensing images within a certain element or element set range can be accurately and quickly realized.

The invention provides an automatic registration method of vector data and remote sensing images, comprising the following steps:

Step 1. Based on the vector data, use the scan line algorithm to extract the ternary matrix A;

Step 2. Extract the grayscale matrix G based on the remote sensing image, and use the Canny operator to obtain the edge matrix E of the grayscale matrix G;

Step 3, use the edge matrix E to traverse the three-valued matrix A to generate the normalized edge registration factor matrix F'_side;

Step 4. Use the grayscale matrix G to traverse the three-valued matrix A to generate a normalized grayscale registration factor matrix F'_gray;

Step 5, based on the characteristics of the registration factor, perform automatic registration of the vector data and the remote sensing image;

Step 6: Extract remote sensing image data within a specified range of vector elements.

Step 1 specifically includes:

1.1. Extract the minimum outer rectangle of the vector data, and convert it into a binarized grid matrix according to formula (1) Y={y(i,j)|i=0,...,m-1; j=0 ,...,n-1}, where m is the height of the outer rectangle of the vector data, and n is the width of the outer rectangle of the vector data;

1.2. Create a Boolean flag matrix F={f(i,j)|i=0,...,m-1; j=0,...,n-1} to judge whether it is a vector element edge;

1.3. For each row scan line L _t ={(t,j)|j=0,...,n-1}(t∈[0,m-1]), use formula (2) to F for assignment;

1.4. Identify the boundaries and interiors of vector elements according to the marker matrix F; create a matrix A={a(i,j)|i=0,...,m-1; j=0,...,n- 1}, the initial default value is 0, and the assignment is performed according to formula (3) to obtain a three-valued matrix A after the assignment, wherein the edge point identifier of the vector data is 2, the interior point identifier is 1, and the background point identifier is 0;

Step 2 specifically includes:

2.1. Extract grayscale matrix based on public remote sensing images: read remote sensing image data to matrix C={c(i,j)|i=0,...,p-1; j=0,...,q- 1}, it is processed as a grayscale matrix G={g(i,j)|i=0,...,p-1; j=0,...,q-1} according to formula (4). ;

g(i,j)=0.299*c _r (i,j)+0.587*c _g (i,j)+0.114*c _b (i,j) (4)

Among them, p is the height of remote sensing image data, q is the width of remote sensing image data, and satisfy the conditions (p>m) and (q>n); _{cr (i,j), c g (} i, j), c _b ( i, j) represent the R, G, and B values of the remote sensing image C at point (i, j), respectively;

2.2. Use a Gaussian filter to smooth the grayscale matrix G;

a) Use formula (5) to calculate the Gaussian convolution kernel G'={ ^g '(x,y)|x=-k, . ..,k; y=-k,...,k};

b) Convolve the convolution kernel G' with the grayscale matrix G of the remote sensing image to obtain the smoothed image matrix S={s(i,j)|i=0,...,p-1; j=0, ...,q-1};

2.3. Calculate the gradient magnitude and direction; use formulas (6), (7), (8) to calculate the magnitude and direction of the gradient of the smoothing matrix S;

θ(i,j)=arctan( _Px (i,j)/ _Py (i,j)) (8)

Among them, P _x and P _y are the gradient operators of the image in the x and y directions, respectively, and P _x (i,j) and P _y (i, j) are the gradient operators at the point (i, j) and the smoothing matrix. Product, arctan represents the tangent function, M(i,j) is the amplitude of the smoothing matrix S at (i,j), θ(i,j) is the direction of the smoothing matrix S at (i,j);

2.4. Non-maximum suppression of gradient amplitude: According to formula (9), the gradient matrix after non-maximum suppression is obtained Grad={grad(i,j)|i=0,...,p-1 ; j=0,...,q-1};

Wherein _front M represents the gradient magnitude of the previous point along the gradient direction (i, j), and _rear M represents the gradient magnitude of the next point along the gradient direction (i, j);

2.5. Edge detection and connection according to the double-threshold method: establish a high threshold δ _high and a low threshold δ _low , and the point (i, j) that satisfies the condition (10) can be determined as an edge point, and these points are connected to get the final result. The edge image matrix E={e(i,j)|i=0,...,p-1; j=0,...,q-1};

g(i,j)> _δhigh or(g(i,j)< _δhigh and g(i,j)> _δlow and flag(i,j)=true) (10)

Step 3 specifically includes:

3.1. Set the point at the upper left corner of the remote sensing image as the origin, and use the edge image matrix E and the ternary matrix A to obtain the point pair of the relative displacement deviation between the upper left corner of the user's own vector data and the origin of the remote sensing image when the user's own vector data is in different positions. Set D={(d _x ,d _y )|0≤d _x ≤pm-1; 0≤d _y ≤qn-1}, according to formula (11), calculate any deviation (d _x ,d _y in set D) ) corresponding result matrix T _dxdy ={t(i,j)|i=0,...,m-1; j=0,...,n-1};

3.2. According to formula (12), calculate the sum f of the elements in the result matrix T corresponding to the deviation;

3.3. The elements corresponding to the elements in the set D and the resulting matrix F _edge = {f _edge (x, y) | x = 0, ..., pm-1; y = 0, ..., qn-1 } is the edge registration factor matrix;

3.4. Normalize the obtained edge registration factor matrix F _edge according to formula (13) to obtain the normalized edge registration factor matrix F' _edge = {f' _edge (d _x , _dy )|d _x =0,...,pm-1; _{dy =} 0,...,qn-1};

Among them, the f- _{side max} represents the maximum value of the edge registration factor matrix, and the f- _{side min} represents the minimum value of the edge registration factor matrix.

Step 4 specifically includes:

4.1. Using the grayscale image matrix G and the three-valued matrix A, for any deviation {(d _x , _dy )} in the set D, calculate the corresponding result matrix T'={t'( i,j)|i=0,...,m-1; j=0,...,n-1};

4.2. Calculate the variance f' of non-zero values in the corresponding result matrix according to formulas (15) and (16);

where R is the number of non-zero values in the result matrix T';

4.3. The resulting matrix F _gray = {f _gray (x, y)|x = 0,..., pm-1; y = 0,..., qn-1} composed of the variance corresponding to each element in the set D is the grayscale registration factor matrix;

4.4. Normalize the obtained grayscale registration factor matrix F _gray according to formula (17) to obtain a normalized grayscale registration factor matrix F' _gray = {f' _gray (d _x , _dy )| d _x =0,...,pm-1; _dy =0,...,qn-1};

Among them, f _{gray max} represents the maximum value of the gray registration factor matrix, and f _{gray min} represents the minimum value of the gray registration factor matrix.

Step 5 specifically includes:

5.1. From the characteristics of the registration factor, it can be known that the larger the edge registration factor of a certain deviation position is and the smaller the grayscale registration factor is, the more matched the vector data at the deviation position is with the remote sensing image. Following this principle, formula (18) is used. Create a comprehensive registration factor matrix F = _{{fjudgment (} i,j)|i=0,...,pm-1; j=0,...,qn-1};

5.2. Traverse the discriminant value matrix _F , and obtain the deviation value (i ₀ , j ₀ ) corresponding to the minimum value, which is the optimal deviation value of the registration of the vector data and the remote sensing image.

Step 6 specifically includes:

Generate the result matrix R={r(x,y)|x=0,...,m-1; y=0,...,n-1} according to formula (19); the image corresponding to this matrix is a vector Remote sensing imagery images within the range of data elements or feature sets, and the rest are black background;

Beneficial effect: Compared with the prior art, the present invention proposes a grayscale registration factor matrix obtained by comprehensively utilizing remote sensing grayscale images and an edge registration factor matrix obtained by combining with the Canny operator, using the characteristics of both It is a method to solve and optimize the deviation between remote sensing images and own vector data, and finally obtain remote sensing images within a certain range of vector elements. This method mainly has the following characteristics:

1) Make full use of the large position offset between the own vector data and remote sensing images, and the shape and angle deformation can be ignored;

2) The automatic registration of vector elements and remote sensing images is realized by comprehensively utilizing the consistency between remote sensing image edge images and vector elements and the correlation of objects within a specific range.

Description of drawings

Fig. 1 is the flow chart of the method of the present invention;

Figure 2 is a schematic diagram of the deviation between remote sensing images and vector data;

Fig. 3 is the remote sensing image data used in the embodiment;

Fig. 4 is the vector data used in the embodiment;

5 is a schematic diagram of the initial relative position of remote sensing image data and vector data;

6 is a schematic diagram of a ternary matrix of vector data;

Figure 7 is a grayscale image of a remote sensing image;

Figure 8 is a remote sensing image edge image;

FIG. 9 is a schematic diagram of the corresponding point product of the ternary matrix and the edge image when the deviation is (0,0);

10 is a schematic diagram of an edge registration factor matrix;

Figure 11 is a schematic diagram of the corresponding dot product part of a three-value matrix and a grayscale image when the deviation is (0,0);

12 is a schematic diagram of a grayscale registration factor matrix;

13 is a schematic diagram of a comprehensive registration factor matrix;

Figure 14 is a schematic diagram of the relative position after the registration of the remote sensing image and the vector data;

Figure 15 shows the results of extracting remote sensing images within the range of vector elements.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

In this example, the base map of remote sensing images of Jiangsu Province in 2016 released by Tiantu-Jiangsu is used as the remote sensing image data (Fig. 3); the user-owned vector data is a part of the park element resources in Wuxi City (Fig. 4). The relative positions of the two starting points are shown in Figure 5. Among them, the size of the remote sensing map is 708*571 pixels, and the size of the user-owned vector data is 612*512 pixels.

Step 1. Extract the ternary matrix of vector data;

1.1. Extract the minimum outer rectangle of the vector data, and convert it into a binarized grid matrix according to formula (1) Y={y(i,j)|i=0,...,m-1; j=0 ,...,n-1}, as shown in Figure 4, in this embodiment, m=708, n=571;

1.2. Create a Boolean flag matrix F={f(i,j)|i=0,...,m-1; j=0,...,n-1} to judge whether it is a vector element edge;

1.3. For each row scan line L _t ={(t,j)|j=0,...,n-1}(t∈[0,m-1]), use formula (2) to F for assignment;

1.4. According to the mark matrix F, mark the boundary and interior of the vector elements respectively. Create a matrix A={a(i,j)|i=0,...,m-1; j=0,...,n-1}, the initial value is 0, and the assignment is obtained according to formula (3). A three-valued matrix A, in which the edge point of the vector data is identified as 2, the interior point is identified as 1, and the background point is identified as 0. In this embodiment, the obtained three-valued matrix is shown in FIG. 6 . For display purposes, each value of the matrix is enlarged by 255/2 times as the pixel value of the image.

Step 2. Extract the grayscale matrix and the edge matrix based on the remote sensing image;

2.1. Extract the grayscale matrix based on the public remote sensing image, and read the remote sensing image data to the matrix C={c(i,j)|i=0,...,p-1; j=0,...,q- 1}, it is processed as a grayscale matrix G={g(i,j)|i=0,...,p-1; j=0,...,q-1} according to formula (4). , as shown in FIG. 7 , in this embodiment, p=612, q=512;

2.2. Use a Gaussian filter to smooth the grayscale matrix G;

a) Calculate the Gaussian convolution kernel G'={ ^g '(x,y)|x=-k,.. .,k; y=-k,...,k}, in this embodiment, k=1, σ ² =1;

b) Convolve the convolution kernel G' with the grayscale matrix G of the remote sensing image to obtain the smoothed image matrix S={s(i,j)|i=0,...,p-1; j=0, ...,q-1};

2.3. Calculate the gradient magnitude and direction, and use formulas (6), (7), and (8) to calculate the magnitude and direction of the gradient of the smoothing matrix S;

2.4. Non-maximum suppression of gradient amplitude. According to formula (9), the gradient matrix after non-maximum suppression is obtained Grad={grad(i,j)|i=0,...,p-1; j=0,...,q-1} ;

2.5. Edge detection and connection are carried out according to the double threshold method. A high threshold delta _high =150 and a low threshold delta _low =50 were established. The point (i, j) that satisfies the condition (10) can be determined as an edge point. Connect these points to obtain the final edge image matrix E={e(i,j)|i=0,...,p-1; j=0,...,q-1}, as shown in Figure 8 shown.

Step 3, generating a normalized edge registration factor matrix;

3.1. Set the point in the upper left corner of the remote sensing image as the origin, and use the edge image matrix E and the ternary matrix A to obtain the deviation of the relative displacement between the upper left corner of the remote sensing image and the origin of the remote sensing image when the user's own vector data is in different positions For the set D ₌ {( _dx ,dy)| _0≤dx≤pm -1; _0≤dy≤qn -1}. Calculate the result matrix T={t(i,j)|i=0,...,m-1; j=0,...,n-1} according to formula (11) , some of its elements are shown in Figure 9;

3.2. According to formula (12), calculate the corresponding matrix elements and f=550;

3.3. The result matrix F _edge corresponding to set D={f _edge (x,y)|x=0,...,pm-1; y=0,...,qn-1} is the edge registration factor matrix.

3.4. Normalize the obtained _edge registration factor matrix F according to formula (13) to obtain the normalized edge registration factor matrix F' _side ={f' _side (d _x , _dy )|d _x = 0,...,pm-1; _dy= 0,...,qn-1}, as shown in Figure 10.

Step 4, generating a normalized grayscale registration factor matrix;

4.1. Using the grayscale image matrix G and the three-valued matrix A, for the deviation (0,0), calculate the corresponding result matrix T'={t'(i,j)|i=0,. ..,m-1; j=0,...,n-1}, some of its elements are shown in Figure 11;

4.2. Calculate the variance f'=1555 of non-zero values in the corresponding result matrix according to formulas (15) and (16);

4.3. The variance matrix F _gray corresponding to the set D = {f _gray (x, y) | x = 0, ..., pm-1; y = 0, ..., qn-1} is gray registration factor matrix.

4.4. Normalize the obtained grayscale registration factor matrix F _gray according to formula (17) to obtain the normalized grayscale registration factor matrix F' _gray = {f' _gray (d _x , _dy )|d _x =0,...,pm-1; _dy= 0,...,qn-1}, as shown in FIG. 12 .

Step 5, based on the characteristics of the registration factor matrix, perform automatic registration of the vector data and the remote sensing image;

5.1. From the characteristics of the registration factor, it can be known that the larger the edge registration factor of a certain deviation position is and the smaller the gray registration factor is, the more matched the vector data at the deviation position is with the remote sensing image. Following this principle, formula (18) is used to create a comprehensive registration factor matrix _F = {f = (d _x , _dy ) | d _{x =} 0,..., pm-1; _dy = 0, .. .,qn-1}, as shown in Figure 13;

5.2. Traverse the discriminant value matrix F to obtain the deviation value ( _65,47 ) corresponding to the minimum value. This is the optimal deviation value of vector data and remote sensing image registration, and the registration result is shown in Figure 14.

Step 6: Extract remote sensing image data within a specified range of vector elements.

The resulting image matrix R={r(x,y)|x=0,...,m-1;y=0,...,n-1} is generated according to formula (19). The image corresponding to this matrix (Fig. 15) is the remote sensing image image corresponding to the vector data elements, and the rest is a black background.

It can be seen from the above embodiments that the method can automatically and accurately extract remote sensing image data within a certain vector range. Compared with the existing registration methods, this method is mainly aimed at the automatic registration between the user's own vector data and the public remote sensing image data. registration processing needs.

In this embodiment, the edge registration factor and the grayscale registration factor are selected to perform registration in a manner of equal weight, which basically meets the needs of extracting remote sensing image images within this range. Different types of vector data have different weight settings for the registration factor. If the vector data represents large homogeneous natural features such as forests and lakes, the grayscale registration factor should take a larger proportion; if it represents residential areas, which are not homogeneous but have clear boundaries For artificial features, the edge registration factor should account for a larger proportion.

Claims (7)

1. An automatic registration method of vector data and remote sensing images is characterized in that: the method comprises the following steps:

step 1, extracting a ternary matrix A by using a scanning line algorithm based on vector data;

step 2, extracting a gray matrix G based on the remote sensing image, and obtaining an edge matrix E of the gray matrix G by using a Canny operator;

step 3, traversing the three-value matrix A by using the edge matrix E to generate a normalized edge registration factor matrix F' _Edge ；

Step 4, traversing the three-value matrix A by using the gray-scale matrix G to generate a normalized gray-scale registration factor matrix F' _{Ash of} ；

Step 5, automatically registering the vector data and the remote sensing image based on the characteristics of the registration factor;

And 6, extracting the remote sensing image data in the specified range of the vector elements.

2. The automatic registration method of vector data and remote sensing images according to claim 1, characterized in that: the step 1 specifically comprises:

1.1, extracting a minimum outsourcing rectangle of the vector data, and converting the minimum outsourcing rectangle into a binary grid matrix Y ═ { Y (i, j) | i ═ 0, ·, m-1 according to a formula (1); j is 0, a., n-1, wherein m is the height of the vector data outsourcing rectangle, and n is the width of the vector data outsourcing rectangle;

1.2, creating a boolean flag matrix F ═ { F (i, j) | i ═ 0. j ═ 0.., n-1} determines whether or not the vector element edge is present;

1.3 scanning the line L for each row _t ＝{(t,j)|j＝0,...,n-1}(t∈[0,m-1]) Assigning the mark matrix F by using a formula (2);

1.4, respectively marking the boundary and the interior of the vector element according to the mark matrix F; creating a matrix a ═ { a (i, j) | i ═ 0. j is 0, the.. and n-1, the initial default value is 0, and assignment is performed according to a formula (3) to obtain an assigned ternary matrix A, wherein the edge point identifier of the vector data is 2, the internal point identifier is 1, and the background point identifier is 0;

3. the automatic registration method of vector data and remote sensing images according to claim 2, characterized in that: the step 2 specifically comprises:

2.1, extracting a gray matrix based on the public remote sensing image: reading remote sensing image data to a matrix C (C (i, j) | i ═ 0., p-1; j-0., q-1} which is processed as a gray matrix G ═ { G (i, j) | i ═ 0., p-1; j ═ 0.., q-1 };

g(i,j)＝0.299*c _r (i,j)+0.587*c _g (i,j)+0.114*c _b (i,j) (4)

wherein, p is the height of the remote sensing image data, q is the width of the remote sensing image data, and the conditions (p > m) and (q > n) are satisfied; c. C _r (i,j)、c _g (i,j)、c _b (i, j) respectively represents R, G, B values of the remote sensing image C at the point (i, j);

2.2, smoothing the gray matrix G by using a Gaussian filter;

a) calculating the user given size (2k +1) × (2k +1) and the variance σ using equation (5) ² A lower gaussian convolution kernel G '═ { G' (x, y) | x ═ k, · k; y ═ k,.., k };

b) convolving the convolution kernel G' with the remote sensing image gray matrix G to obtain a smoothed image matrix S ═ { S (i, j) | i ═ 0., p-1; j ═ 0.., q-1 };

2.3, calculating the amplitude and the direction of the gradient; calculating the amplitude and the direction of the gradient of the smoothing matrix S by using the formulas (6), (7) and (8);

θ(i,j)＝arctan(P _x (i,j)/P _y (i,j)) (8)

wherein P is _x ，P _y Gradient operators, P, in the x, y directions of the image, respectively _x (i,j)，P _y (i, j) is the product of the gradient operator at point (i, j) and the smoothing matrix, arctan represents the tangent function, M (i, j) is the magnitude of the smoothing matrix S at (i, j), θ (i, j) is the direction of the smoothing matrix S at (i, j);

2.4, carrying out non-maximum suppression on the gradient amplitude: obtaining a gradient matrix Grad ═ { Grad (i, j) | i ═ 0.., p-1 after the non-maximum value is suppressed according to the formula (9); j ═ 0.., q-1 };

wherein M is _{Front side} Representing the magnitude of the gradient, M, at a point preceding point (i, j) in the direction of the gradient _{Rear end} Represents the gradient magnitude of a point subsequent to the point (i, j) in the gradient direction;

2.5, carrying out edge detection and connection according to a double-threshold method: establishing a high threshold delta _{Height of} And a low threshold delta _{Is low in} Points (i, j) satisfying the condition (10) can be determined as edge points, and the points are connected to obtain a final edge image matrix E ═ { E (i, j) | i ═ 0. j ═ 0.., q-1 };

g(i,j)＞δ _{height of} or(g(i,j)＜δ _{Height of} and g(i,j)＞δ _{Is low in} and flag(i,j)＝true) (10)

4. The automatic registration method of vector data and remote sensing images according to claim 3, characterized in that: the step 3 specifically includes:

3.1, setting the point at the upper left corner of the remote sensing image as an origin, and obtaining a point pair set D { (D) of the relative displacement deviation between the point at the upper left corner and the origin of the remote sensing image when the user own vector data are at different positions by using the edge image matrix E and the ternary matrix A _x ,d _y )|0≤d _x ≤p-m-1；0≤d _y Q-n-1 ≦ q, any deviation (D) in the set D is calculated according to equation (11) _x ,d _y ) Corresponding result matrix

3.2, calculating the sum f of elements in the result matrix T corresponding to the deviation according to the formula (12);

3.3 element corresponding to each element in the set D and the resulting matrix F formed _Edge ＝{f _Edge (x, y) | x ═ 0.., p-m-1; y is 0,., q-n-1, which is the edge registration factor matrix;

3.4, obtaining an edge registration factor matrix F _Edge Normalizing according to a formula (13) to obtain a normalized edge registration factor matrix F' _Edge ＝{f' _Edge (d _x ,d _y )|d _x ＝0,...,p-m-1；d _y ＝0,...,q-n-1}；

Wherein f is _{Side max} Representing the maximum value of the edge registration factor matrix, f _{While min} Represents the minimum of the edge registration factor matrix.

5. The automatic registration method of vector data and remote sensing images according to claim 4, characterized in that: the step 4 specifically includes:

4.1 Using the grayscale image matrix G and the ternary matrix A, any deviation { (D) in the set D _x ,d _y ) A corresponding result matrix T '═ { T' (i, j) | i ═ 0., m-1 is calculated according to formula (14); j ═ 0.., n-1 };

4.2, calculating the variance f' of the non-zero value in the corresponding result matrix according to the formulas (15) and (16);

wherein R is the number of non-zero values in the result matrix T';

4.3 result matrix F consisting of variances corresponding to elements in set D _{Ash of} ＝{f _{Ash of} (x, y) | x ═ 0.., p-m-1; y is 0, the q-n-1 is a gray level registration factor matrix;

4.4, obtaining a gray level registration factor matrix F _{Ash of} Normalizing according to the formula (17) to obtain a normalized gray level registration factor matrix F' _{Ash of} ＝{f' _{Ash of} (d _x ,d _y )|d _x ＝0,...,p-m-1；d _y ＝0,...,q-n-1}；

Wherein f is _{Ash max} Representing the maximum value of the gray scale registration factor matrix, f _{Lime min} Representing the minimum value of the gray scale registration factor matrix.

6. The automatic registration method of vector data and remote sensing images according to claim 5, characterized in that: the step 5 specifically includes:

5.1, according to the characteristic of the registration factor, if the edge registration factor of a deviation position is larger and the gray level registration factor is smaller, the vector data at the deviation position is matched with the remote sensing image more, and the comprehensive registration factor matrix F is created by using a formula (18) according to the principle _Judgment ＝{f _Judgment (i,j)|i＝0,...,p-m-1；j＝0,...,q-n-1}；

5.2 traversing the discriminating momentArray F _Judgment Obtaining the deviation value (i) corresponding to the minimum value ₀ ,j ₀ ) And the vector data is the optimal deviation value of the registration of the vector data and the remote sensing image.

7. The automatic registration method of vector data and remote sensing images according to claim 6, characterized in that: the step 6 specifically includes:

generating a result matrix R ═ { R (x, y) | x ═ 0., m-1 according to equation (19); y is 0,., n-1 }; the image corresponding to the matrix is a remote sensing image in a vector data element or element set range, and the rest part is a black background;

Images