Remote sensing image registration method and device, storage medium and terminal equipment

By extracting and matching the advanced texture boundary lines of remote sensing images and using primary matrix constraints for image registration, the problem of low registration accuracy in weak texture areas is solved by traditional methods, and efficient and accurate image registration is achieved.

CN118967754APending Publication Date: 2024-11-15BEIJING DATA INTELLIGENCE INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411058656.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-02
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

Traditional remote sensing image registration methods are not effective when dealing with weak texture areas. It is difficult to extract feature points based on feature points. Registration based on image grayscale values ​​is not sensitive to the light receiving environment, and registration time based on mutual information is long. And it may fall into local extreme values, resulting in low registration accuracy and complex operation of single texture and weak texture images.

Method used

By obtaining the reference image and the image to be registered, the primary matrix is ​​calculated, cut into sub-blocks, the advanced texture coefficient is extracted, the boundary lines of weak texture areas are identified and matched, and the primary matrix constraints are used to filter the matching pairs, and the transformation matrix is ​​calculated for registration.

Benefits of technology

Accurate and fast image registration in single textured, weak texture areas are achieved, improving the accuracy and reliability of registration and reducing operational complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure BDA0004977147100000031
    Figure BDA0004977147100000031
  • Figure BDA0004977147100000034
    Figure BDA0004977147100000034
  • Figure BDA0004977147100000035
    Figure BDA0004977147100000035
Patent Text Reader

Abstract

The invention provides a remote sensing image registration method and device, a storage medium and terminal equipment. The method comprises the following steps: S1, acquiring a reference image and an image to be registered, forming an image pair and calculating a primary matrix; s2, acquiring a plurality of reference image sub-blocks and a plurality of to-be-registered image sub-blocks; s3, extracting weak texture areas in the reference image and the to-be-registered image in the image pair; s4, identifying primary boundary lines of weak texture areas in the reference image and the to-be-registered image respectively, and performing matching to obtain a primary boundary line matching pair set; s5, screening an advanced boundary line matching pair set; s6, obtaining a transformation matrix of the weak texture area in the reference image and the weak texture area in the to-be-registered image; and S7, solving the positions of the pixel points in the weak texture area in the reference image in the weak texture area in the to-be-registered image by using the transformation matrix. According to the registration method provided by the invention, accurate and rapid registration can also be realized under the condition that weak textures exist in the image. The problem that in the prior art, weak texture areas of images cannot be matched or are poor in matching performance is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of image registration, and in particular to a remote sensing image registration method and device, a storage medium and a terminal device. Background Art

[0002] Remote sensing image registration is an important step in remote sensing image processing. It aims to align remote sensing images acquired at different times, from different sensors or from different perspectives so that they can match each other in geographic space. In remote sensing applications, accurate image registration is of vital importance for subsequent image analysis, change detection, map making, etc. The current mainstream automatic image registration methods mainly include registration based on feature points, registration based on image grayscale values, and registration methods based on mutual information.

[0003] However, traditional remote sensing image registration methods often do not work well when dealing with weak texture areas. For remote sensing images with single texture or weak texture, it is difficult to accurately extract feature points based on feature point registration; registration based on image grayscale values ​​requires that the grayscale values ​​of the two images must be consistent, and the registration accuracy is not high for images affected by the environment such as lighting; registration based on mutual information takes a long time to register, and may fall into local extreme values, making it impossible to obtain accurate registration results. Weak texture areas refer to areas in the image that lack obvious texture features or structural information, such as water bodies and snow. At present, many images with single texture or weak texture still use manual registration. The success rate and accuracy of manual registration are relatively high, but it increases the burden on operators and the registration speed is relatively slow.

[0004] In order to solve the above problems, the matching of weak texture and single texture areas in images often faces challenges such as low registration accuracy, slow speed or complex operation. Therefore, it is of great practical significance and application value to develop a registration method that can efficiently and accurately process single texture and weak texture remote sensing images. Summary of the invention

[0005] In order to solve the above technical problems, the first purpose of the present invention is to provide a remote sensing image registration method, which can achieve accurate and fast registration even in scenes with single texture or weak texture in the image.

[0006] A remote sensing image registration method, the method comprising:

[0007] S1 acquires a reference image and an image to be registered and forms an image pair; and calculates the primary matrix of the reference image and the image to be registered in the image pair;

[0008] S2 respectively cuts the reference image and the image to be registered in the image pair to obtain a plurality of reference image sub-blocks and a plurality of image sub-blocks to be registered;

[0009] S3 sequentially extracts weak texture areas in the reference image and the image to be registered in the image pair based on the high-level reference texture coefficients of each reference image sub-block and the high-level to-be-registered texture coefficients of each to-be-registered image sub-block;

[0010] S4 respectively identifies primary boundary lines of weak texture areas in the reference image and the image to be registered, and matches them based on the primary boundary lines of the two to obtain a primary boundary line matching pair set;

[0011] S5 screens out an advanced boundary line matching pair set from the primary boundary line matching pair set based on the constraints of the primary matrix;

[0012] S6 obtains the transformation matrix of the weak texture area in the reference image and the weak texture area in the image to be registered based on the advanced boundary line matching pair set;

[0013] S7 uses the transformation to obtain the position of the pixel points in the weak texture area of ​​the image to be registered in the weak texture area of ​​the reference image, and completes the registration of the image to be registered and the reference image in the weak texture area.

[0014] Specifically, step S3 includes:

[0015] S31 respectively calculates the primary texture coefficients of each reference image sub-block and each to-be-registered image sub-block and performs normalization processing to obtain a high-level reference texture coefficient set and a high-level to-be-registered texture coefficient set;

[0016] S32, according to a preset texture threshold, selecting multiple reference image sub-blocks corresponding to advanced reference texture coefficients less than the preset texture threshold from the advanced reference texture coefficient set as weak texture areas in the reference image;

[0017] S33 selects, according to a preset texture threshold, a plurality of image sub-blocks to be registered corresponding to high-level texture coefficients to be registered that are smaller than the preset texture threshold from the high-level texture coefficient set to be registered as weak texture areas in the image to be registered.

[0018] Specifically, the method for normalizing the primary texture coefficients in step S31 is:

[0019] First, calculate the variance of each image sub-block, and the calculation formula is:

[0020]

[0021] in, is the variance of the image sub-block, n is the number of the image sub-blocks, x i For any pixel point, is the pixel average value of the intercepted image sub-block; the image sub-block includes the reference image sub-block and each image sub-block to be registered;

[0022] Then, based on the variance of the image sub-block, its primary texture coefficient is calculated using the following formula:

[0023]

[0024] Among them, t x is the primary texture coefficient, t min =0.5,ε Var =0.00005;

[0025] The primary texture coefficients are normalized to obtain the advanced texture coefficients. The normalization formula is:

[0026]

[0027] Where T is the high-level texture coefficient, t x is a primary texture coefficient, max is a maximum value among the multiple primary texture coefficients, and min is a minimum value among the multiple primary texture coefficients.

[0028] Specifically, step S4 also includes:

[0029] S41 performs binarization processing on the reference image and the image to be registered according to a preset texture threshold value to obtain a binary image of the reference image and a binary image of the image to be registered;

[0030] S42 extracts a primary boundary line set of a weak texture region in the reference image and a primary boundary line set of a weak texture region in the image to be registered, respectively, based on the binary image of the reference image and the binary image of the image to be registered;

[0031] S43 respectively obtains the intersection points of any two primary boundary lines in the reference image and the intersection points of any two primary boundary lines in the image to be registered;

[0032] S44 matches multiple primary boundary lines in the primary boundary line set of the weak texture area in the reference image with multiple primary boundary lines in the primary boundary line set of the weak texture area in the image to be registered through line segment descriptors to obtain a primary boundary line matching pair set.

[0033] Specifically, step S5 further includes: determining in turn whether each primary boundary line matching pair in the primary boundary line matching pair set satisfies the constraints of the primary matrix, and selecting multiple primary boundary line matching pairs that satisfy the constraints of the primary matrix as the advanced boundary line matching pair set.

[0034] Specifically, step S2 is as follows: randomly select a pixel point in the reference image, take the pixel point as the center, and cut out a reference image sub-block according to a preset pixel area; repeat the above operation until the reference image is cut into multiple reference image sub-blocks; and cut the image to be registered into multiple image sub-blocks to be registered using the above operation.

[0035] The second object of the present invention is to provide a remote sensing image registration device, the device comprising:

[0036] A primary matrix acquisition module is used to calculate the primary matrix of the reference image and the image to be registered;

[0037] A pixel cropping module, used for cropping the reference image and the image to be registered into a plurality of reference image sub-blocks and image sub-blocks to be registered respectively;

[0038] A weak texture extraction module is used to sequentially extract weak texture areas in the reference image and the image to be registered in the image pair;

[0039] A boundary line matching module is used to identify and match high-level boundary lines of weak texture areas in the reference image and high-level boundary lines of weak texture areas in the image to be registered, and obtain a high-level boundary line matching pair set;

[0040] The transformation matrix module is used to calculate the transformation matrix between the weak texture area in the reference image and the weak texture area in the image to be registered, and to find the position of the pixel points in the weak texture area in the image to be registered in the weak texture area in the reference image based on the transformation matrix.

[0041] The third object of the present invention is to provide a computer-readable storage medium, which stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in the above-mentioned remote sensing image registration method.

[0042] Another object of the present invention is a terminal device, comprising: a processor, a memory and a communication bus; the memory stores a computer-readable program executable by the processor;

[0043] The communication bus realizes the connection and communication between the processor and the memory;

[0044] When the processor executes the computer-readable program, the steps in the above-mentioned remote sensing image registration method are implemented.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] (1) The present invention cuts the reference image and the image to be registered into multiple image sub-blocks respectively, and regards multiple reference image sub-blocks and multiple image sub-blocks to be registered corresponding to high-level texture coefficients less than the texture threshold as weak texture areas of the reference image and the image to be registered respectively; the present invention can more accurately identify and extract weak texture areas by refining image pixel units, which is convenient for the subsequent effective registration of weak texture areas;

[0047] (2) The present invention uses the boundary lines of weak texture areas for matching, and the boundary lines contain more significant structural information, so more reliable registration results can be obtained in weak texture areas;

[0048] (3) The present invention introduces the constraint conditions of the primary matrix to screen the boundary line matching pairs, thereby further improving the accuracy and reliability of the matching between the reference image and the image to be registered in the weak texture area. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0050] Figure 1 A technical flow chart of a remote sensing image registration method in an embodiment of the present invention;

[0051] Figure 2 A technical solution diagram of a remote sensing image registration method in an embodiment of the present invention;

[0052] Figure 3 Schematic diagram of a method for calculating a transformation matrix in an embodiment of the present invention;

[0053] Figure 4 A technical scheme diagram of a remote sensing image registration device in an embodiment of the present invention;

[0054] Figure 5 The figure is a schematic diagram of a framework of a terminal device in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The following will be combined with the accompanying drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0056] See also Figure 1 and Figure 2 , Figure 1 A technical flow chart of a remote sensing image registration method in an embodiment of the present invention; Figure 2 : is a technical scheme diagram of a remote sensing image registration method in an embodiment of the present invention; the method comprises:

[0057] S1 acquires a reference image and an image to be registered and forms an image pair; and calculates a primary matrix of the reference image and the image to be registered in the image pair.

[0058] In the embodiment of the present invention, the feature points extracted by SIFT are used for matching, and a primary matrix between the reference image and the image to be registered is calculated.

[0059] S2 cuts the reference image and the image to be registered in the image pair into a plurality of reference image sub-blocks and a plurality of image sub-blocks to be registered respectively.

[0060] In an embodiment of the present invention, step S2 is specifically as follows: randomly select a pixel point in the reference image, take the pixel point as the center, and cut out a reference image sub-block according to a preset pixel area; repeat the above operation until the reference image is cut into multiple reference image sub-blocks; and cut the image to be registered into multiple image sub-blocks to be registered using the above operation.

[0061] Specifically, first randomly select a pixel in the reference image; with the pixel as the center, select a number of pixels with equal rows and columns to form a reference image sub-block of N×N size; then repeat randomly selecting a pixel, again with the pixel as the center, select a number of rows and columns of pixels from top to bottom and left to right to cut out the reference image sub-block, until the reference image is divided into multiple reference image sub-blocks of N×N size; it should be noted that each pixel can only be taken as the center point once. The image to be registered is cut into multiple N×N image sub-blocks to be registered with the same operation as above. This solution can cut out reference image sub-blocks and image sub-blocks to be registered of size 5×5, and can also cut out image sub-blocks of size 3×3, 7×7, and 9×9.

[0062] S3 sequentially extracts weak texture areas in the reference image and the image to be registered in the image pair based on the high-level reference texture coefficients of each reference image sub-block and the high-level to-be-registered texture coefficients of each to-be-registered image sub-block.

[0063] In this embodiment of the present invention, step S3 includes:

[0064] S31 calculates primary texture coefficients of each reference image sub-block and each to-be-registered image sub-block respectively and performs normalization processing to obtain a high-level reference texture coefficient set and a high-level to-be-registered texture coefficient set.

[0065] In the embodiment of the present invention, the method for normalizing the primary texture coefficients in step S31 is:

[0066] First, calculate the variance of each image sub-block, and the calculation formula is:

[0067]

[0068] in, is the variance of the image sub-block, n is the number of the image sub-blocks, x is an arbitrary pixel point, is the pixel average value of the intercepted image sub-block; the image sub-block includes the reference image sub-block and each image sub-block to be registered;

[0069] Then, based on the variance of the image sub-block, its primary texture coefficient is calculated using the following formula:

[0070]

[0071] Among them, t x is the primary texture coefficient, t min =0.5,ε Var =0.00005;

[0072] The primary texture coefficients are normalized to obtain the advanced texture coefficients. The normalization formula is:

[0073]

[0074] Where T is the high-level texture coefficient, t x is a primary texture coefficient, max is a maximum value among the multiple primary texture coefficients, and min is a minimum value among the multiple primary texture coefficients.

[0075] S32, according to a preset texture threshold, selecting multiple reference image sub-blocks corresponding to advanced reference texture coefficients less than the preset texture threshold from the advanced reference texture coefficient set as weak texture areas in the reference image;

[0076] S33 selects, according to a preset texture threshold, a plurality of image sub-blocks to be registered corresponding to high-level texture coefficients to be registered that are smaller than the preset texture threshold from the high-level texture coefficient set to be registered as weak texture areas in the image to be registered.

[0077] In the embodiment of the present invention, all primary texture coefficients are normalized based on the normalization formula to obtain the high-level reference texture coefficient of each reference image sub-block and the high-level texture coefficient to be registered of the image sub-block to be registered, and each high-level reference texture coefficient and high-level texture coefficient to be registered are compared with the preset texture threshold. The smaller the high-level texture coefficient is, the weaker the texture of the image sub-block is. The continuous area composed of the reference image sub-blocks corresponding to the high-level reference texture coefficients less than the texture threshold in the high-level reference texture coefficients is the weak texture area in the reference image; the continuous area composed of the image sub-blocks corresponding to the high-level texture coefficients less than the texture threshold in the high-level texture coefficients to be registered is the weak texture area in the image to be registered. It should be noted that the threshold can be set arbitrarily, and the size of the continuous area can be controlled by adjusting the set threshold; for example, if the texture threshold is set to 0.5, the area with a high-level texture coefficient less than 0.5 is regarded as a weak texture area, or if the threshold is set to 0.3, the area with a high-level texture coefficient less than 0.3 is a weak texture area; by adjusting the texture threshold, the texture areas of different degrees in the image can be controlled as weak texture areas.

[0078] S4 identifies primary boundary lines of weak texture areas in the reference image and the image to be registered, respectively, and performs matching based on the primary boundary lines of the two to obtain a primary boundary line matching pair set.

[0079] In this embodiment of the present invention, step S4 further includes:

[0080] S41 performs binarization processing on the reference image and the image to be registered according to a preset texture threshold value to obtain a binary image of the reference image and a binary image of the image to be registered;

[0081] S42 extracts a primary boundary line set of a weak texture region in the reference image and a primary boundary line set of a weak texture region in the image to be registered, respectively, based on the binary image of the reference image and the binary image of the image to be registered;

[0082] S43 respectively obtains the intersection points of any two primary boundary lines in the reference image and the intersection points of any two primary boundary lines in the image to be registered;

[0083] S44 matches multiple primary boundary lines in the primary boundary line set of the weak texture area in the reference image with multiple primary boundary lines in the primary boundary line set of the weak texture area in the image to be registered through line segment descriptors to obtain a primary boundary line matching pair set.

[0084] In the embodiment of the present invention, the reference image sub-blocks belonging to the weak texture area in the reference image are marked as 1, and the other areas are marked as 0, so as to obtain a binary image of the reference image; the image sub-blocks to be registered belonging to the weak texture area in the image to be registered are marked as 1, and the other areas are marked as 0, so as to obtain a binary image of the image to be registered. For the obtained binary images, EDlines straight line detection is performed respectively to extract the primary boundary line set of the weak texture area in the reference image and the primary boundary line set of the weak texture area in the image to be registered.

[0085] S5 selects an advanced boundary line matching pair set from the primary boundary line matching pair set based on the constraints of the primary matrix.

[0086] In the embodiment of the present invention, step S5 further includes: determining in turn whether each primary boundary line matching pair in the primary boundary line matching pair set satisfies the constraints of the primary matrix, and selecting multiple primary boundary line matching pairs that satisfy the constraints of the primary matrix as the advanced boundary line matching pair set.

[0087] In the embodiment of the present invention, the intersection of any two primary boundary lines in the reference image and the intersection of any two primary boundary lines in the image to be registered are obtained respectively; and the primary boundary lines extracted from the reference image and the image to be registered are matched by the LBD line descriptor. If the two primary boundary lines in the reference image are matched with the two primary boundary lines in the image to be registered respectively, then their intersection points are also matching points. Since the matching points need to satisfy the constraints of the primary matrix, that is, the matching points are on each other's epipolar lines, this property is further used to reversely verify the correctness of the line matching, screen out some erroneous line matching, and obtain a high-level boundary line matching pair set.

[0088] S6 obtains the transformation matrix of the weak texture area in the reference image and the weak texture area in the image to be registered based on the advanced boundary line matching pair set.

[0089] S7 uses the transformation to obtain the position of the pixel points in the weak texture area of ​​the image to be registered in the weak texture area of ​​the reference image, and completes the registration of the image to be registered and the reference image in the weak texture area.

[0090] See also Figure 3 , Figure 3 Schematic diagram of the calculation method of the transformation matrix in an embodiment of the present invention; through the above steps, multiple high-level boundary line matching pairs describing the weak texture plane are obtained, and the transformation matrix of the weak texture plane in the reference image and the image to be registered is calculated by selecting the high-level boundary line matching pairs that match each other, and the matching relationship of any point between the image to be registered and the reference image is obtained through the transformation matrix. Figure 3 The upper image is the reference image, and the lower image is the image to be registered). The quadrilateral box represents the boundary of the weak texture area, and the dotted line in the upper image represents the epipolar line of the point in the lower image in the upper image. Figure 3As shown, I1 matches I1', I2 matches I2', and it has been determined that p2 matches p2' as well. In order to solve the transformation matrix, three pairs of matching points are required, namely Figure 3 Next, we will introduce in detail how to calculate the above three points in Figure 3 Corresponding points in the upper middle image: When the two high-level boundary lines I and I' in a pair of high-level boundary line matching pairs between the reference image and the image to be registered and the primary matrix between the reference image and the image to be registered are known, the matching point p' of any point p on I in the other image can be solved. Since p is on the straight line I, the corresponding point p' of p must be on the straight line I' of the other image, and must also be on the polar line of p. The two straight lines determine a unique intersection point, so the intersection point p' of I' and the polar line of p in the other image is the corresponding point of p. In summary, the correspondence between the points on the straight line can be determined through the correspondence between the straight lines under two perspectives and the primary matrix.

[0091] The present invention also provides a remote sensing image registration device 100. Figure 4 , Figure 4 This is a technical scheme diagram of a remote sensing image registration device in an embodiment of the present invention. The device includes:

[0092] A primary matrix acquisition module 101 is used to calculate the primary matrix of the reference image and the image to be registered;

[0093] A pixel cropping module 102, used to crop the reference image and the image to be registered into a plurality of reference image sub-blocks and image sub-blocks to be registered respectively;

[0094] A weak texture extraction module 103 is used to sequentially extract weak texture areas in the reference image and the image to be registered in the image pair;

[0095] The boundary line matching module 104 is used to identify and match the high-level boundary lines of the weak texture area in the reference image and the high-level boundary lines of the weak texture area in the image to be registered, so as to obtain a high-level boundary line matching pair set;

[0096] The transformation matrix module 105 is used to calculate the transformation matrix between the weak texture area in the reference image and the weak texture area in the image to be registered, and to obtain the position of the pixel points in the weak texture area in the image to be registered in the weak texture area in the reference image based on the transformation matrix.

[0097] In an embodiment of the present invention, a computer-readable storage medium is also provided, which stores one or more programs. The one or more programs can be executed by one or more processors to implement the steps in the above-mentioned remote sensing image registration method.

[0098] See also Figure 5 , Figure 5 The present invention also provides a terminal device 200, including: a processor 201, a memory 202, a communication bus 203 and a display 204; the memory 202 stores a computer-readable program that can be executed by the processor;

[0099] The communication bus 203 realizes the connection and communication between the processor and the memory;

[0100] When the processor 201 executes the computer-readable program, the steps in the above-mentioned remote sensing image registration method are implemented.

[0101] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) that contain computer-usable program code.

[0102] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0103] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0104] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0105] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A remote sensing image registration method, It is characterized in that The method comprises: S1 acquires a reference image and an image to be registered and forms an image pair; and calculates the primary matrix of the reference image and the image to be registered in the image pair; S2 respectively cuts the reference image and the image to be registered in the image pair to obtain a plurality of reference image sub-blocks and a plurality of image sub-blocks to be registered; S3 sequentially extracts weak texture areas in the reference image and the image to be registered in the image pair based on the high-level reference texture coefficients of each reference image sub-block and the high-level to-be-registered texture coefficients of each to-be-registered image sub-block; S4 respectively identifies primary boundary lines of weak texture areas in the reference image and the image to be registered, and matches them based on the primary boundary lines of the two to obtain a primary boundary line matching pair set; S5 screens out an advanced boundary line matching pair set from the primary boundary line matching pair set based on the constraints of the primary matrix; S6 obtains the transformation matrix of the weak texture area in the reference image and the weak texture area in the image to be registered based on the advanced boundary line matching pair set; S7 uses the transformation to obtain the position of the pixel points in the weak texture area of ​​the image to be registered in the weak texture area of ​​the reference image, and completes the registration of the image to be registered and the reference image in the weak texture area.

2. The method according to claim 1, It is characterized in that Step S3 includes: S31 respectively calculates the primary texture coefficients of each reference image sub-block and each to-be-registered image sub-block and performs normalization processing to obtain a high-level reference texture coefficient set and a high-level to-be-registered texture coefficient set; S32, according to a preset texture threshold, selecting multiple reference image sub-blocks corresponding to advanced reference texture coefficients less than the preset texture threshold from the advanced reference texture coefficient set as weak texture areas in the reference image; S33 selects, according to a preset texture threshold, a plurality of image sub-blocks to be registered corresponding to high-level texture coefficients to be registered that are smaller than the preset texture threshold from the high-level texture coefficient set to be registered as weak texture areas in the image to be registered.

3. The method according to claim 2, It is characterized in that The method for normalizing the primary texture coefficients in step S31 is: First, calculate the variance of each image sub-block, and the calculation formula is: in, is the variance of the image sub-block, n is the number of the image sub-blocks, x i For any pixel point, is the pixel average value of the intercepted image sub-block; the image sub-block includes the reference image sub-block and each image sub-block to be registered; Then, based on the variance of the image sub-block, its primary texture coefficient is calculated using the following formula: Among them, t x is the primary texture coefficient, t min =0.5,ε Var =0.00005; The primary texture coefficients are normalized to obtain the advanced texture coefficients. The normalization formula is: Where T is the high-level texture coefficient, t x is a primary texture coefficient, max is a maximum value among the multiple primary texture coefficients, and min is a minimum value among the multiple primary texture coefficients.

4. The method according to claim 2, It is characterized in that Step S4 also includes: S41 performs binarization processing on the reference image and the image to be registered according to a preset texture threshold value to obtain a binary image of the reference image and a binary image of the image to be registered; S42 extracts a primary boundary line set of a weak texture region in the reference image and a primary boundary line set of a weak texture region in the image to be registered, respectively, based on the binary image of the reference image and the binary image of the image to be registered; S43 respectively obtains the intersection points of any two primary boundary lines in the reference image and the intersection points of any two primary boundary lines in the image to be registered; S44 matches multiple primary boundary lines in the primary boundary line set of the weak texture area in the reference image with multiple primary boundary lines in the primary boundary line set of the weak texture area in the image to be registered through line segment descriptors to obtain a primary boundary line matching pair set.

5. The method according to claim 1, It is characterized in that Step S5 also includes: determining in turn whether each primary boundary line matching pair in the primary boundary line matching pair set satisfies the constraints of the primary matrix, and selecting multiple primary boundary line matching pairs that satisfy the constraints of the primary matrix as the advanced boundary line matching pair set.

6. The method according to claim 1, It is characterized in that Step S2 is specifically as follows: randomly select a pixel point in the reference image, take the pixel point as the center, and cut out a reference image sub-block according to a preset pixel area; repeat the above operation until the reference image is cut into multiple reference image sub-blocks; and cut the image to be registered into multiple image sub-blocks to be registered using the above operation.

7. A remote sensing image registration device, It is characterized in that include: A primary matrix acquisition module is used to calculate the primary matrix of the reference image and the image to be registered; A pixel cropping module, used for cropping the reference image and the image to be registered into a plurality of reference image sub-blocks and image sub-blocks to be registered respectively; A weak texture extraction module is used to sequentially extract weak texture areas in the reference image and the image to be registered in the image pair; A boundary line matching module is used to identify and match high-level boundary lines of weak texture areas in the reference image and high-level boundary lines of weak texture areas in the image to be registered, and obtain a high-level boundary line matching pair set; The transformation matrix module is used to calculate the transformation matrix between the weak texture area in the reference image and the weak texture area in the image to be registered, and to find the position of the pixel points in the weak texture area in the image to be registered in the weak texture area in the reference image based on the transformation matrix.

8. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps in a remote sensing image registration method as described in any one of claims 1-6.

9. A terminal device, It is characterized in that include: Processor, memory and communication bus; The memory stores a computer-readable program executable by the processor; The communication bus realizes the connection and communication between the processor and the memory; When the processor executes the computer-readable program, the steps in the remote sensing image registration method described in any one of claims 1 to 6 are implemented.