Image matching coarse error correction method, system and device and storage medium

Abstract

The invention provides an image matching coarse error correction method, system and device and a storage medium, and the method comprises the steps: obtaining a stereopair connection point containing a certain proportion of matching coarse error through image matching, and calculating an initial correlation coefficient of the stereopair connection point; determining sequence matching windows used by the stereo image pair connection points and the number of the sequence matching windows; under the sequence matching window, a weighted least square rank coordinate correction model and a phase correlation rank coordinate correction model are constructed according to the stereo image pair connection points, and corrected rank coordinates and correlation coefficients are obtained; comparing the correlation coefficient obtained under the sequence matching window of the connection point of the same stereopair with the initial correlation coefficient, and selecting the corrected row and column coordinates obtained under the corresponding sequence matching window as the corrected row and column coordinates of the connection point of the stereopair when the correlation coefficient is maximum; therefore, the coarse error of the connection point obtained by image matching is corrected.

Description

Technical Field

[0001] The present invention belongs to the technical field of photogrammetry, and in particular relates to a method, system, device and storage medium for correcting gross errors in image matching. Background Art

[0002] Different fields have different application requirements and technical details for image matching. At present, in the field of photogrammetry, image matching has been widely used in aspects such as digital surface model generation, three-dimensional scene reconstruction, geometric processing and change detection. The image matching described is based on the requirement of obtaining correct connection point pairs. By analyzing the consistency and similarity of related information such as features and structures between different images, the positions of the same ground point on different images are identified and located, providing observation values ​​for subsequent steps such as direct geometric positioning of images and bundle adjustment. The connection points obtained by image matching contain a certain proportion of mismatches. If there are mismatches, the subsequent direct geometric positioning, bundle adjustment and other steps will be affected in the form of gross errors in observation values. Therefore, the accuracy of image matching is one of the most critical parameters. Summary of the Invention

[0003] In response to the above-mentioned deficiencies or improvement needs of the prior art, the present invention proposes a method, system, device and storage medium for correcting coarse errors in image matching to correct coarse errors in connection points obtained by image matching.

[0004] To achieve the above object, according to a first aspect of the present invention, a method for correcting gross errors in image matching is provided, comprising the following steps: S1. Obtaining stereo image pair connection points containing a certain proportion of matching coarse errors through image matching, and calculating the initial correlation coefficients of the stereo image pair connection points; S2. Determine the sequence matching window and number used for the stereo pair connection points; S3. Under the sequence matching window, construct a weighted least squares row and column coordinate correction model according to the stereo image pair connection points to obtain corrected row and column coordinates and correlation coefficients; S4. For the same stereo image pair connection point, construct a phase-correlated row and column coordinate correction model under the sequence matching window to obtain corrected row and column coordinates and correlation coefficients; S5. For the same stereo image pair connection point, compare the correlation coefficient obtained in the sequence matching window with the initial correlation coefficient, and select the corrected row and column coordinates obtained in the sequence matching window corresponding to the maximum correlation coefficient as the corrected row and column coordinates of the stereo image pair connection point.

[0005] According to the above method, the S2 specifically includes: S201, determining the number of pixels between each stereo image pair connection point and the four boundaries of the left and right images, and determining the minimum number of pixels; S202, performing contour detection and calculating the pixels of the minimum circumscribed circle; S203: When the connection point of the stereo image pair is within the minimum circumscribed circle, the smaller value between the minimum circumscribed circle pixel radius and the minimum number of pixels is taken as the maximum matching window radius. S204 , the sequence matching window is increased from a matching window radius of 1 pixel and decreased to a maximum matching window radius, and the increasing and decreasing step values ​​are the same; and then the sequence matching window and the number are determined.

[0006] According to the above method, in S204, when the minimum number of pixels exceeds the given number of windows, the number of sequence matching windows is determined by the given number of windows.

[0007] According to the above method, the S3 specifically includes: S301, determining initial values ​​of deformation parameters of the left and right images; S302, bilinearly resampling the left and right images according to the initial value; S303: Differentiation in row and column directions, formulating error equations while taking into account both geometric distortion and radiation distortion, and solving correlation coefficients. The loop iteration ends when the correlation coefficient is less than a preset limit. S304: Perform weighted averaging on the row and column coordinates to obtain corrected row and column coordinates and correlation coefficients.

[0008] According to the above method, in S302, the point to be interpolated is first located, then the four nearest discrete data points are found, and the horizontal interpolation factor and the vertical interpolation factor between the coordinates of the point to be interpolated and the four nearest discrete data points are calculated. Then, a weighted average is performed on the four nearest discrete data points, and the weight is calculated according to the interpolation factor.

[0009] According to the above method, the S4 specifically includes: S401, after determining the sequence matching windows and the number of the stereo image pair connection points, opening left and right window images of the same size on the left and right images respectively according to the window size, and eliminating edge effects on the left and right window images; S402, performing low-pass filtering on the left and right window images that eliminate edge effects; S403, converting the left and right window images after low-pass filtering from the time domain to the frequency domain through Fourier transform; S404, solving the normalized cross-power spectrum matrix in the frequency domain, extracting phase information by calculating the normalized cross-power spectrum matrix between the two left and right window images; S405, inversely transforming the left and right window images back to the time domain through inverse Fourier transform; S406, determining the maximum window offset parameter and setting the loop iteration condition; S407: When the iteration is completed, the corrected row and column coordinates and correlation coefficients are obtained.

[0010] According to a second aspect of the present invention, there is provided an image matching coarse error correction system, comprising: A module for obtaining stereo image pair connection points with coarse errors is used to obtain stereo image pair connection points with a certain proportion of coarse matching errors through image matching, and calculate the initial correlation coefficients of the stereo image pair connection points; A sequence matching window determination module, used to determine the sequence matching windows and the number of sequence matching windows used for the connection points of the stereo image pair; A weighted least squares row and column coordinate correction module is used to construct a weighted least squares row and column coordinate correction model according to the stereo image pair connection points under the sequence matching window to obtain corrected row and column coordinates and correlation coefficients; A phase-correlated row and column coordinate correction module is used to construct a phase-correlated row and column coordinate correction model for the same stereo image pair connection point under the sequence matching window to obtain corrected row and column coordinates and correlation coefficients; The correction result output module is used to compare the correlation coefficient obtained under the sequence matching window and the initial correlation coefficient for the same stereo image pair connection point, and select the corrected row and column coordinates obtained under the corresponding sequence matching when the correlation coefficient window is the largest as the corrected row and column coordinates of the stereo image pair connection point.

[0011] According to the above system, the sequence matching window determination module is specifically used for: Determine the number of pixels between each stereo pair connection point and the four boundaries of the left and right images, and determine the minimum number of pixels; Perform contour detection and calculate the pixels of the minimum circumscribed circle; When the connection point of the stereo image pair is in the minimum circumscribed circle, the smaller value between the minimum circumscribed circle pixel radius and the minimum number of pixels is taken as the maximum matching window radius; The sequence matching window is increased from a matching window radius of 1 pixel and decreased to a maximum matching window radius, and the increasing and decreasing step values ​​are the same; thereby determining the sequence matching window and its number.

[0012] According to a third aspect of the present invention, a device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0013] According to a fourth aspect of the present invention, there is provided a storage medium having a computer program stored thereon, wherein the computer program implements the steps of the above method when executed by a processor.

[0014] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art: The present invention sets an adaptive sequence matching window, constructs a weighted least squares row and column coordinate correction model and a phase correlation row and column coordinate correction model respectively, compares the correlation coefficient obtained under the adaptive sequence matching window with the initial correlation coefficient, selects the corrected row and column coordinates obtained under the sequence matching window when the correlation coefficient is maximum as the corrected row and column coordinates of this coarse matching point pair, and then corrects the coarse errors of the connection points obtained by image matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the method provided by an embodiment of the present invention.

[0016] Figure 2 Schematic diagram of an adaptive sequence window provided by an embodiment of the present invention.

[0017] Figure 3 It is a schematic diagram of a weighted least squares row-column coordinate correction model provided by an embodiment of the present invention.

[0018] Figure 4 Schematic diagram of a phase-correlated row and column coordinate correction model provided by an embodiment of the present invention.

[0019] Figure 5 This is a scene diagram of experimental data located in a certain area provided by an embodiment of the present invention.

[0020] Figure 6 2 is a schematic diagram for comparing symmetric transfer errors provided by an embodiment of the present invention.

[0021] Figure 7 It is a schematic diagram of the system structure provided by an embodiment of the present invention.

[0022] Figure 8 It is a structural diagram of a device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0024] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0025] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise specifically defined. In the present invention, unless otherwise specified or limited, the terms "installed," "connected," "connect," "fixed," etc. should be understood in a broad sense. For example, they can refer to fixed connection, detachable connection, or integration; mechanical connection, electrical connection; direct connection, or indirect connection through an intermediate medium; internal communication between two components, or interaction between two components, unless otherwise specified. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0026] In the present invention, unless otherwise expressly specified or limited, when a first feature is "above" or "below" a second feature, it may mean that the first and second features are in direct contact, or that the first and second features are in indirect contact through an intermediary. Furthermore, when a first feature is "above," "above," or "above" a second feature, it may mean that the first feature is directly above or diagonally above the second feature, or simply means that the first feature is at a higher level than the second feature. When a first feature is "below," "below," or "below" a second feature, it may mean that the first feature is directly below or diagonally below the second feature, or simply means that the first feature is at a lower level than the second feature.

[0027] According to the first aspect of the present invention, a method for coarse error correction based on comprehensive weighted least squares and phase correlation matching of an adaptive sequence window is provided. Figure 1 As shown, the following steps are included: S1. Obtain stereo image pair connection points containing a certain proportion of matching coarse errors through image matching, and calculate the initial correlation coefficients of the stereo image pair connection points.

[0028] Specifically, in some embodiments, an image matching method such as SIFT is first used to obtain stereo image pair connection points containing a certain proportion of rough matching errors. Taking SIFT as an example, the following implementation methods are included: S101, left and right image feature detection, first, the left and right images are scale-space generated, and the scale-space extreme points are detected, then the extreme points are accurately located, and the key point direction parameters are determined; S102, feature description of the left and right images, obtaining target feature point sets of the left and right images through key point descriptor generation; S103, matching feature points of the left and right images, completing the matching of feature points of the left and right images based on similarity measures such as Euclidean distance, and performing preliminary elimination of rough matching based on the RANSAC algorithm, thereby obtaining stereo image pair connection points containing a certain proportion of rough matching errors; S104 , calculating the initial correlation coefficient of the connection points of the stereo image pair in S103 .

[0029] S2. Determine the sequence matching window and quantity used for the stereo pair connection points.

[0030] The present invention introduces a sequence window (i.e., an adaptive sequence window) of gradually changing size for each pair of homonymous points in the rough matching, such as Figure 2 As shown, it is mainly divided into the following steps: S201 , determining the number of pixels between each stereo image pair connection point and the four boundaries of the left and right images, and determining the minimum number of pixels.

[0031] S202 , performing contour detection based on the Canny algorithm and calculating the pixels of the minimum circumscribed circle; other algorithms may also be used to accomplish this.

[0032] S203. When the connection point of the stereo image pair is in the minimum circumscribed circle, the smaller value between the minimum circumscribed circle pixel radius and the minimum number of pixels is taken as the maximum matching window radius. That is, the minimum circumscribed circle pixel radius and the minimum number of pixels are compared. When the minimum circumscribed circle radius is greater than the minimum number of pixels, the minimum number of pixels is used as the maximum matching window radius; when the minimum circumscribed circle radius is less than the minimum number of pixels, the minimum circumscribed circle radius is used as the maximum matching window radius.

[0033] S204 , the sequence matching window is increased from a matching window radius of 1 pixel and decreased to a maximum matching window radius, and the increasing and decreasing step values ​​are the same; and then the sequence matching window and the number are determined.

[0034] In some embodiments, the step value is 2 pixels, that is, the sequence window increases and decreases in both directions from a matching window radius of 1 pixel to a maximum matching window radius, each time increasing and decreasing by 2 pixels.

[0035] To balance computational efficiency, when the determined minimum number of pixels exceeds the given number of windows, the number of sequence matching windows is determined by the given number of windows.

[0036] S3, under the sequence matching window, constructing a weighted least squares row and column coordinate correction model according to the stereo image pair connection points to obtain the corrected row and column coordinates and correlation coefficients; Figure 3 As shown, it specifically includes the following steps: S301, determine the initial values ​​of the left and right image deformation parameters. In some embodiments, the initial values ​​of the left and right image deformation parameters are set to h 0=0, h 1 = 1, a 0= x 1- x 2 ，a 1=1, a 2=0, b 0= y 1- y 2, b 1=0, b 2=1.

[0037] in h 0 and h 1 is the radiation distortion parameter, a 0 -a 2 and b 0 -b 2 is the geometric distortion parameter, ( x 1, y 1) and ( x 2, y 2) The image plane coordinates of the center positions of the left and right image windows.

[0038] S302: bilinearly resample the left and right images according to the initial values.

[0039] For bilinear interpolation, we usually first locate the point to be interpolated (x, y), where x and y are the horizontal and vertical coordinates of the point to be interpolated, respectively. Then, we find the four nearest discrete data points and calculate the relative positions of x and y to these four nearest discrete data points, i.e., the horizontal interpolation factor and the vertical interpolation factor. Then, we take a weighted average of the four nearest discrete data points, with the weights calculated according to the interpolation factor.

[0040] S303, perform row and column direction differences, and formulate an error equation while taking both geometric distortion and radiation distortion into account. The error equation can be expressed as:

[0041] in and is the random error at the interpolation point. The correlation coefficient is calculated, and the loop iteration ends when the correlation coefficient is less than a preset limit. The size of the limit determines the accuracy of the optimal point.

[0042] S304: Perform weighted averaging on the row and column coordinates to obtain corrected row and column coordinates and correlation coefficients.

[0043] S4. For the same stereo image pair connection point, a phase-correlated row and column coordinate correction model is constructed under the sequence matching window to obtain the corrected row and column coordinates and correlation coefficients; Figure 4 As shown, the method mainly includes the following sub-steps: S401, after determining the sequence matching windows and the number of the stereo image pair connection points, respectively opening left and right window images of the same size on the left and right images according to the window size, and eliminating edge effects on the left and right window images; S402, performing low-pass filtering on the left and right window images to eliminate edge effects; S403, converting the left and right window images from the time domain to the frequency domain through Fourier transform; S404 , solving the normalized cross-power spectrum matrix in the frequency domain, and extracting phase information by calculating the normalized cross-power spectrum matrix between the two images.

[0044] S405 , inversely transform the left and right window images back to the time domain through inverse Fourier transform.

[0045] S406: Determine the maximum window offset parameter and set loop iteration conditions.

[0046] S407: When the iteration is completed, the corrected row and column coordinates and correlation coefficients are obtained.

[0047] S5. For the same stereo pair connection point, compare all correlation coefficients obtained within the sequential matching window with the initial correlation coefficient, and select the corrected row and column coordinates obtained within the sequential matching window corresponding to the maximum correlation coefficient as the corrected row and column coordinates of the coarse matching point pair (i.e., the stereo pair connection point). All correlation coefficients include the correlation coefficients obtained in S3 and S4.

[0048] Finally, the corrected row and column coordinate values ​​of all stereo image pair connection points are obtained.

[0049] It should be understood that the order of execution of the steps in the above embodiments does not necessarily imply a specific order of execution. The execution order of each process should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of this application. In particular, S3 and S4 are executed in parallel.

[0050] The present invention will be further described below by taking a linear array satellite image of a certain area as an example.

[0051] Figure 5 For a linear array satellite image of a certain area, the experimental results are compared from two aspects: symmetric transfer error and direct positioning based on matching point pairs.

[0052] Symmetric transfer error refers to the sum of the squared Euclidean distances between the matching points measured or matched on the left and right images of a stereo pair and the theoretical matching points obtained through homography transformation on the right and left images, respectively. It is a commonly used and important metric for measuring matching accuracy. In this example, 30 pairs of sub-pixel matching points were manually measured, and homography transformation parameters were calculated using a cubic polynomial. The symmetric transfer error was then calculated based on these transformation model parameters. Figure 6 The symmetric transfer error diagram of the matching point pairs is given. In the comparison of the symmetric transfer error, the correction strategy combining weighted least squares and phase correlation under the sequence window achieves the best result.

[0053] According to the second aspect of the present invention, a comprehensive weighted least squares and phase correlation matching coarse error correction system of an adaptive sequence window is provided, such as Figure 7 Shown, including: A module for obtaining stereo image pair connection points with coarse errors is used to obtain stereo image pair connection points with a certain proportion of coarse matching errors through image matching, and calculate the initial correlation coefficients of the stereo image pair connection points; A sequence matching window determination module, used to determine the sequence matching windows and the number of sequence matching windows used for the connection points of the stereo image pair; A weighted least squares row and column coordinate correction module is used to construct a weighted least squares row and column coordinate correction model according to the stereo image pair connection points under the sequence matching window to obtain corrected row and column coordinates and correlation coefficients; A phase-correlated row and column coordinate correction module is used to construct a phase-correlated row and column coordinate correction model for the same stereo image pair connection point under the sequence matching window to obtain corrected row and column coordinates and correlation coefficients; The correction result output module is used to compare the sizes of all correlation coefficients obtained under the sequence matching window and the initial correlation coefficient for the same stereo image pair connection point, and select the corrected row and column coordinates obtained under the corresponding sequence matching window when the correlation coefficient is the largest as the corrected row and column coordinates of this rough matching point pair.

[0054] Wherein, the sequence matching window determination module is specifically used for: Determine the number of pixels between each stereo pair connection point and the four boundaries of the left and right images, and determine the minimum number of pixels; Perform contour detection and calculate the pixels of the minimum circumscribed circle; When the connection point of the stereo image pair is in the minimum circumscribed circle, the smaller value between the minimum circumscribed circle pixel radius and the minimum number of pixels is taken as the maximum matching window radius; The sequence matching window is increased from a matching window radius of 1 pixel and decreased to a maximum matching window radius, and the increasing and decreasing step values ​​are the same; thereby determining the sequence matching window and its number.

[0055] Based on the same inventive concept, the present invention also provides a structure of a computer device, such as Figure 8 As shown, the computer device 20 of this embodiment includes at least but not limited to: a memory 21 and a processor 22 that can communicate with each other via a system bus. It should be noted that Figure 8 Computer device 20 is shown only with components 21 - 22 , but it should be understood that implementing all of the illustrated components is not a requirement, and greater or fewer components may alternatively be implemented.

[0056] In this embodiment, memory 21 (i.e., a readable storage medium) includes flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), and programmable read-only memory (PROM). Memory 21 may also be an external storage device of computer device 20, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash memory card, etc. Of course, memory 21 may also include both internal storage units of computer device 20 and external storage devices. In this embodiment, memory 21 is typically used to store the operating system and various application software installed on computer device 20, such as the program code of the visual recognition device for disordered sorting of cylindrical bars in the method embodiment. Furthermore, memory 21 may also be used to temporarily store various types of data that have been output or are about to be output.

[0057] In some embodiments, the processor 22 can be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 22 is generally used to control the overall operation of the computer device 20. In this embodiment, the processor 22 is used to execute program code stored in the memory 21 or process data, for example, to execute the visual recognition device for disordered sorting of cylindrical bars to implement the visual recognition method for disordered sorting of cylindrical bars in the method embodiment.

[0058] Based on the same inventive concept, the present application also provides a computer-readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory), a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a disk, an optical disk, a server, an App store, etc., on which a computer program is stored, and when the program is executed by a processor, the corresponding function is implemented. The computer-readable storage medium of this embodiment is used to store a system for coarse error correction using a comprehensive weighted least squares and phase-correlated matching method for an adaptive sequence window, and when executed by a processor, implements a method for coarse error correction using a comprehensive weighted least squares and phase-correlated matching method for an adaptive sequence window.

[0059] It should be pointed out that, according to the needs of implementation, the various steps / components described in this application can be split into more steps / components, or two or more steps / components or partial operations of steps / components can be combined into new steps / components to achieve the purpose of the present invention.

[0060] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for correcting gross errors in image matching, characterized by: The following steps are involved: S1. Obtaining stereo image pair connection points containing a certain proportion of matching coarse errors through image matching, and calculating the initial correlation coefficients of the stereo image pair connection points; S2. Determine the sequence matching window and number used for the stereo pair connection points; S3. Under the sequence matching window, construct a weighted least squares row and column coordinate correction model according to the stereo image pair connection points to obtain corrected row and column coordinates and correlation coefficients; S4. For the same stereo image pair connection point, construct a phase-correlated row and column coordinate correction model under the sequence matching window to obtain corrected row and column coordinates and correlation coefficients; S5. For the same stereo image pair connection point, compare all correlation coefficients obtained in the sequence matching window with the initial correlation coefficient, and select the corrected row and column coordinates obtained in the sequence matching window corresponding to the maximum correlation coefficient as the corrected row and column coordinates of the stereo image pair connection point.

2. The method according to claim 1, wherein: The S2 specifically includes: S201, determining the number of pixels between each stereo image pair connection point and the four boundaries of the left and right images, and determining the minimum number of pixels; S202, performing contour detection and calculating the pixels of the minimum circumscribed circle; S203: When the connection point of the stereo image pair is within the minimum circumscribed circle, the smaller value between the minimum circumscribed circle pixel radius and the minimum number of pixels is taken as the maximum matching window radius. S204 , the sequence matching window is increased from a matching window radius of 1 pixel and decreased to a maximum matching window radius, and the increasing and decreasing step values ​​are the same; and then the sequence matching window and the number are determined.

3. The method according to claim 2, wherein: In the above S204, when the minimum number of pixels exceeds the given number of windows, the number of sequence matching windows is determined by the given number of windows.

4. The method according to claim 1, wherein: The S3 specifically includes: S301, determining initial values ​​of deformation parameters of the left and right images; S302, bilinearly resampling the left and right images according to the initial values; S303: Differentiation in row and column directions, formulating error equations while taking into account both geometric distortion and radiation distortion, and solving correlation coefficients. The loop iteration ends when the correlation coefficient is less than a preset limit. S304: Perform weighted averaging on the row and column coordinates to obtain corrected row and column coordinates and correlation coefficients.

5. The method according to claim 4, characterized in that: In the above-mentioned S302, the point to be interpolated is first located, then the four nearest discrete data points are found, and the horizontal interpolation factor and the vertical interpolation factor between the coordinates of the point to be interpolated and the four nearest discrete data points are calculated. Then, a weighted average is performed on the four nearest discrete data points, and the weight is calculated according to the interpolation factor.

6. The method according to claim 1, wherein: The S4 specifically includes: S401, after determining the sequence matching windows and the number of the stereo image pair connection points, opening left and right window images of the same size on the left and right images respectively according to the window size, and eliminating edge effects on the left and right window images; S402, performing low-pass filtering on the left and right window images to eliminate edge effects; S403, converting the left and right window images after low-pass filtering from the time domain to the frequency domain through Fourier transform; S404, solving the normalized cross-power spectrum matrix in the frequency domain, extracting phase information by calculating the normalized cross-power spectrum matrix between the two left and right window images; S405, inversely transforming the left and right window images back to the time domain through inverse Fourier transform; S406, determining the maximum window offset parameter and setting the loop iteration condition; S407: When the iteration is completed, the corrected row and column coordinates and correlation coefficients are obtained.

7. An image matching gross error correction system, characterized by: include: A module for obtaining stereo image pair connection points with coarse errors is used to obtain stereo image pair connection points with a certain proportion of coarse matching errors through image matching, and calculate the initial correlation coefficients of the stereo image pair connection points; A sequence matching window determination module, used to determine the sequence matching windows and the number of sequence matching windows used for the connection points of the stereo image pair; A weighted least squares row and column coordinate correction module is used to construct a weighted least squares row and column coordinate correction model according to the stereo image pair connection points under the sequence matching window to obtain corrected row and column coordinates and correlation coefficients; A phase-correlated row and column coordinate correction module is used to construct a phase-correlated row and column coordinate correction model for the same stereo image pair connection point under the sequence matching window to obtain corrected row and column coordinates and correlation coefficients; The correction result output module is used to compare the correlation coefficient obtained under the sequence matching window and the initial correlation coefficient for the same stereo image pair connection point, and select the corrected row and column coordinates obtained under the corresponding sequence matching window when the correlation coefficient is the largest as the corrected row and column coordinates of the stereo image pair connection point.

8. The image matching coarse error correction system according to claim 7, characterized in that: The sequence matching window determination module is specifically used for: Determine the number of pixels between each stereo pair connection point and the four boundaries of the left and right images, and determine the minimum number of pixels; Perform contour detection and calculate the pixels of the minimum circumscribed circle; When the connection point of the stereo image pair is in the minimum circumscribed circle, the smaller value between the minimum circumscribed circle pixel radius and the minimum number of pixels is taken as the maximum matching window radius; The sequence matching window is increased from a matching window radius of 1 pixel and decreased to a maximum matching window radius, and the increasing and decreasing step values ​​are the same; thereby determining the sequence matching window and its number.

9. A device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images