Oblique image matching method, device, computer equipment and readable storage medium

Abstract

The present invention provides an oblique image matching method, device, computer equipment and readable storage medium, the oblique image matching method comprising: using a reference image and a camera tilt angle, and filtering out multiple related images intersecting with a reference image from multiple oblique images through a preset screening algorithm; calculating the rotation angle of the related image relative to the reference image through a preset rotation angle algorithm, and performing rotation processing on the related image at the rotation angle; performing a preset feature point extraction algorithm on the related image after the rotation processing to obtain a corresponding related feature point set; matching the related feature point set with the reference feature point set of the reference image to obtain a pair of feature points with the same name and having the same features; determining the spatial position relationship between the corresponding related image and the reference image through the pair of feature points with the same name. The oblique image matching method of the present invention can improve the efficiency of the overall oblique image matching and improve the robustness of the oblique image matching.

Description

Technical Field

[0001] The present invention relates to the field of computer vision, and in particular to an inclined image matching method, device, computer equipment and readable storage medium. Background Art

[0002] Existing drones are generally equipped with five-lens cameras. The oblique images obtained by the five lenses have both affine transformation and rotation angle. However, in the existing oblique image matching process, it is impossible to know the rotation angle of the oblique image and select relevant oblique images for matching. Therefore, the overall oblique image matching has the problems of low efficiency and low robustness. Summary of the invention

[0003] In view of the above problems, the present invention provides an oblique image matching method, apparatus, computer device and readable storage medium to improve the efficiency of the overall oblique image matching and improve the robustness of the oblique image matching.

[0004] In order to achieve the above object, the present invention adopts the following technical solution:

[0005] An oblique image matching method, comprising:

[0006] Using the reference image and the camera tilt angle, a plurality of related images intersecting with the reference image are screened out from the plurality of tilted images through a preset screening algorithm;

[0007] Calculating the rotation angle of the related image relative to the reference image by a preset rotation angle algorithm, and performing rotation processing of the related image by the rotation angle;

[0008] Processing the rotated related images with a preset feature point extraction algorithm to obtain a corresponding related feature point set;

[0009] Matching the relevant feature point set with the reference feature point set of the reference image to obtain feature point pairs with the same name and the same features;

[0010] The spatial position relationship between the corresponding related image and the reference image is determined through the pair of feature points with the same name.

[0011] Preferably, in the oblique image matching method, the step of using the reference image and the camera tilt angle to select a plurality of related images intersecting with the reference image from a plurality of oblique images through a preset screening algorithm comprises:

[0012] Calculating the exterior orientation elements of the oblique image according to the camera tilt angle and the initial exterior orientation elements of the downward-looking image;

[0013] Calculating the coverage of the oblique image according to the exterior orientation elements;

[0014] The oblique image whose coverage range intersects with the range of the reference image is determined as the relevant image.

[0015] Preferably, in the oblique image matching method, the preset rotation angle algorithm is an ASIFT algorithm, and the step of calculating the rotation angle of the related image relative to the reference image by using the preset rotation angle algorithm includes:

[0016] Performing downsampling processing on the reference image and the related images at a preset ratio;

[0017] After downsampling, extracting initial matching feature points from the reference image and the related image using the ASIFT algorithm;

[0018] Using the RANSAC algorithm to remove gross error points in the initial matching feature points to obtain matching feature points;

[0019] The rotation angle of the related image relative to the reference image is calculated according to the matching feature points.

[0020] Preferably, in the oblique image matching method, the preset feature point extraction algorithm is a Hessian-Affine algorithm.

[0021] Preferably, in the oblique image matching method, the matching of the relevant feature point set with the reference feature point set of the reference image comprises:

[0022] By performing a matching process of a preset number of related feature point sets and a reference feature point set in parallel using a KD-Tree structure, a plurality of groups of feature point pairs with the same name and the same features are obtained.

[0023] Preferably, the oblique image matching method further comprises:

[0024] After obtaining the feature point pairs with the same name, the RANSAC algorithm is used to eliminate the mismatched point pairs in the feature point pairs with the same name.

[0025] Preferably, the oblique image matching method further comprises:

[0026] The reprojection error of the pair of feature points with the same name is calculated, and the reprojection error is used as a confidence evaluation of the spatial position relationship.

[0027] The present invention also provides an inclined image matching device, comprising:

[0028] A related image screening module, used to screen out a plurality of related images intersecting with the reference image from a plurality of oblique images by using a reference image and a camera tilt angle through a preset screening algorithm;

[0029] A rotation angle calculation module, used to calculate the rotation angle of the related image relative to the reference image by a preset rotation angle algorithm, and perform rotation processing of the related image at the rotation angle;

[0030] A feature point extraction module is used to process the relevant image after the rotation processing with a preset feature point extraction algorithm to obtain a corresponding relevant feature point set;

[0031] A feature point matching module is used to match the relevant feature point set with the reference feature point set of the reference image to obtain feature point pairs with the same name and the same features;

[0032] The position relationship determination module is used to determine the spatial position relationship between the corresponding related image and the reference image through the pair of feature points with the same name.

[0033] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program runs on the processor, the oblique image matching method is executed.

[0034] The present invention also provides a readable storage medium storing a computer program, and the computer program executes the oblique image matching method when running on a processor.

[0035] The present invention provides an oblique image matching method, which includes: using a reference image and a camera tilt angle, and filtering out multiple related images intersecting with the reference image from multiple oblique images through a preset screening algorithm; calculating the rotation angle of the related image relative to the reference image through a preset rotation angle algorithm, and performing rotation processing on the related image by the rotation angle; performing a preset feature point extraction algorithm on the related image after the rotation processing to obtain a corresponding related feature point set; matching the related feature point set with the reference feature point set of the reference image to obtain a pair of feature points with the same features; and determining the spatial position relationship between the corresponding related image and the reference image through the pair of feature points with the same name. The oblique image matching method of the present invention uses the reference image and the camera tilt angle to filter out the related image from the oblique image, which can improve the efficiency of the overall oblique image matching, and by calculating the rotation angle between the related image and the reference image to correct the angle of the related image, supports image matching at a large inclination angle, thereby improving the robustness of oblique image matching.

[0036] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solution of the present invention, the following will briefly introduce the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope of protection of the present invention. In each of the drawings, similar components are numbered similarly.

[0038] Figure 1 is a flow chart of an oblique image matching method provided by Embodiment 1 of the present invention;

[0039] Figure 2 is a flow chart of screening related images provided by Embodiment 2 of the present invention;

[0040] Figure 3 is a flow chart of calculating a rotation angle provided by Embodiment 3 of the present invention;

[0041] Figure 4 is a flow chart of an oblique image matching method provided by Embodiment 4 of the present invention;

[0042] Figure 5 is a flow chart of an oblique image matching method provided by Embodiment 5 of the present invention;

[0043] Figure 6 is a flow chart of another oblique image matching method provided by Embodiment 5 of the present invention;

[0044] Figure 7 It is a structural schematic diagram of an inclined image matching device provided in Example 6 of the present invention. DETAILED DESCRIPTION

[0045] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0046] The components of the embodiments of the present invention generally described and shown in the drawings herein may be arranged and designed in a variety of different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0047] Hereinafter, the terms "including", "having" and their cognates, which may be used in various embodiments of the present invention, are intended only to indicate specific features, numbers, steps, operations, elements, components or combinations of the foregoing items, and should not be understood as first excluding the existence of one or more other features, numbers, steps, operations, elements, components or combinations of the foregoing items or adding the possibility of one or more features, numbers, steps, operations, elements, components or combinations of the foregoing items.

[0048] Furthermore, the terms “first”, “second”, “third”, etc. are merely used for distinguishing descriptions and are not to be understood as indicating or implying relative importance.

[0049] Unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meanings as those generally understood by those skilled in the art to which the various embodiments of the present invention belong. The terms (such as those defined in generally used dictionaries) will be interpreted as having the same meanings as the contextual meanings in the relevant technical field and will not be interpreted as having idealized meanings or overly formal meanings unless clearly defined in the various embodiments of the present invention.

[0050] Example 1

[0051] Figure 1 : is a flowchart of an oblique image matching method provided by Embodiment 1 of the present invention, the method comprising the following steps:

[0052] Step S11: using the reference image and the camera tilt angle, a preset screening algorithm is used to screen out a plurality of related images intersecting with the reference image from a plurality of tilted images.

[0053] In the embodiment of the present invention, oblique image matching is a necessary step in the operation process of drones in the fields of post-image stitching, high-precision stereo mapping, three-dimensional city model reconstruction, image retrieval, etc., and is also a key technology in the field of computer vision. The drone includes multiple viewing angle cameras, such as a lower viewing angle camera, a front viewing angle camera, a rear viewing angle camera, a left viewing angle camera, and a right viewing angle camera, etc. Image stitching refers to the stitching of images between multiple cameras to form a wide-angle image.

[0054] In an embodiment of the present invention, one of the perspective images can be used as a reference image, and the Togo-related images intersecting with the reference image can be screened out from multiple oblique images based on the reference image. The camera tilt angle can be obtained using an accelerometer of a drone, which is not limited here. When screening, the computer device can be provided with an application based on the above-mentioned preset screening algorithm. After obtaining the reference image and the camera tilt angle, they can be input into the application so that the application can screen out multiple related images intersecting with the reference image from multiple oblique images.

[0055] Step S12: calculating the rotation angle of the related image relative to the reference image by a preset rotation angle algorithm, and performing rotation processing of the related image at the rotation angle.

[0056] In the embodiment of the present invention, before matching the positional relationship between the reference image and the related image, it is first necessary to ensure that the two images are on the same horizontal plane, so the related image needs to be rotated to the horizontal plane where the reference image is located. An application based on the above-mentioned preset rotation angle algorithm can be set in the computer device, and the reference image and the related image are input into the application to obtain the above-mentioned rotation angle. By rotating the related image by the rotation angle, the related image and the reference image are finally made to be on the same horizontal plane.

[0057] Step S13: performing a preset feature point extraction algorithm on the relevant image after the rotation process to obtain a corresponding relevant feature point set.

[0058] In the embodiment of the present invention, the preset feature point extraction algorithm is the Hessian-Affine algorithm. That is, an application based on the above Hessian-Affine algorithm can be set in the computer device, and the relevant image after rotation processing can be input into the application to obtain the relevant feature point set corresponding to the relevant image.

[0059] In the embodiment of the present invention, the Hessian-Affine operator detects the initial point through the Hessian matrix in the Gaussian scale space, selects the interest point with the local maximum at multiple scales, and obtains a series of feature points with stable rotation, scale, translation, and brightness changes. The affine transformation threshold is an important parameter for feature point extraction and determines the number of feature points extracted. Due to the large perspective change, rich texture and complex background of oblique images, the use of the same parameter threshold makes the number of feature points extracted from the image vary greatly, and even not in the same order of magnitude, which seriously affects the matching effect. Therefore, a feature point extraction method with an adaptive threshold is proposed to make the number of feature points extracted from the image roughly the same. Set the initial affine transformation parameters to: T=T0, the number of feature points extracted from the image is N, and the parameter transformation step size is step=0.2. The number of feature points extracted from the current film is Ni , when N i <N, change the parameter T=T+step; when N i > N, change the parameter T = T-step; after modifying the parameter, re-extract the feature points until the number of feature points in the image is N i >0.8N&&N i <1.5N meets the requirements.

[0060] Step S14: matching the relevant feature point set with the reference feature point set of the reference image to obtain feature point pairs with the same name and the same features.

[0061] In the embodiment of the present invention, in the process of feature point matching, points with the same features are points with the same name, that is, pairs of feature points with the same name on the reference image and the related image, and the pairs of feature points with the same name can be used as connection points to splice the reference image and the related image. The above matching process can also be implemented using an algorithm or an application. At the beginning of the process step, the Hessian-Affine algorithm can be used to extract a reference feature point set for the reference image. After obtaining the reference feature point set and the related feature point set, they can be input into the matching application to obtain pairs of feature points with the same features and the same name.

[0062] Step S15: determining the spatial position relationship between the corresponding related image and the reference image through the pair of feature points with the same name.

[0063] In the embodiment of the present invention, the spatial positional relationship between the related image and the reference image can be determined according to the positions of the feature point pairs with the same name, that is, the related image and the reference image can be spliced ​​according to the feature point pairs with the same name, thereby forming a wide-angle image for subsequent stereo mapping, three-dimensional city model reconstruction, image retrieval and other field operations.

[0064] In an embodiment of the present invention, relevant images are screened out from oblique images using a reference image and a camera tilt angle, so that the efficiency of the overall oblique image matching can be improved. By calculating the rotation angle between the relevant image and the reference image to correct the angle of the relevant image, image matching at large angles is supported, thereby improving the robustness of oblique image matching.

[0065] Example 2

[0066] Figure 2 This is a flowchart of screening related images provided by Embodiment 2 of the present invention, comprising the following steps:

[0067] Step S21: Calculate the exterior orientation elements of the oblique image according to the camera tilt angle and the initial exterior orientation elements of the downward-looking image.

[0068] In the embodiment of the present invention, the above-mentioned exterior orientation elements include straight line elements, heading angle elements, pitch angle elements and roll angle elements. The exterior orientation elements of the remaining four perspectives can be calculated according to the camera tilt angle and the initial exterior orientation elements of the downward image. The remaining four perspectives include left perspective, right perspective, front perspective and rear perspective.

[0069] Step S22: Calculate the coverage of the oblique image according to the exterior orientation elements.

[0070] In the embodiment of the present invention, the coverage of the current oblique image can be calculated by using the above-mentioned exterior orientation elements and the average ground elevation recorded by the drone. The process of calculating the coverage of the oblique image can be implemented by using an algorithm or an application. For example, an application for calculating the coverage can be set in a computer device, and after obtaining the exterior orientation elements, they can be input into the application to obtain the corresponding coverage.

[0071] Step S23: Determine the oblique image whose coverage range intersects with the range of the reference image as the relevant image.

[0072] In the embodiment of the present invention, after calculating the coverage of the relevant image, the corresponding external box can be solved according to the coverage to determine whether the external box of the oblique image and the external box of the reference image intersect. If there is an intersection, the oblique image can be determined to be a relevant image.

[0073] Example 3

[0074] Figure 3 is a flow chart of calculating a rotation angle provided by Embodiment 3 of the present invention, comprising the following steps:

[0075] Step S31: performing downsampling processing of a preset ratio on the reference image and the related images.

[0076] In the embodiment of the present invention, the preset ratio of the above downsampling process is generally set to 1000 for the longest side, and then downsampling is performed according to the aspect ratio of the original image. Assuming that the width of the image is w and the height is h, the sampling formula is: If w>h, then the sampling ratio is If w<h, the sampling ratio is

[0077] Step S32: After the downsampling process, the ASIFT algorithm is used to extract initial matching feature points from the reference image and the related image.

[0078] In the embodiment of the present invention, when the ASIFT algorithm is used to extract matching feature points, the parameter setting of the ASIFT algorithm may be n=5.

[0079] Step S33: using the RANSAC algorithm to remove gross error points in the initial matching feature points to obtain matching feature points.

[0080] In an embodiment of the present invention, based on the RANSAC algorithm (RANSAC, Random Sample Consensus, random sampling consistency algorithm), a corresponding application can be set up in a computer device. After obtaining the initial matching feature points, the initial matching feature points can be input into the application to obtain the matching feature points after eliminating the gross error points.

[0081] Step S34: Calculating the rotation angle of the related image relative to the reference image according to the matching feature points.

[0082] In the embodiment of the present invention, after obtaining the matching feature points, the transformation matrix can be calculated based on the matching feature points, and the rotation angle can be obtained through the transformation matrix. Assuming that the rotation angle between the images is k, then: g′(x i ,y i )=Mg(x i ,y i ), where M is the rotation matrix.

[0083]

[0084] Calculation of the rotation angle K between images:

[0085] K = arcos(a)

[0086] In an embodiment of the present invention, an application for calculating the above formula may be provided in the computer device, and after the matching feature points are acquired, they may be input into the application to calculate the corresponding rotation angle.

[0087] Example 4

[0088] Figure 4 4 is a flowchart of an oblique image matching method provided by Embodiment 4 of the present invention, the method comprising the following steps:

[0089] Step S41: using the reference image and the camera tilt angle, a preset screening algorithm is used to screen out a plurality of related images intersecting with the reference image from a plurality of tilted images.

[0090] This step is consistent with the above step S11 and will not be described again here.

[0091] Step S42: Calculating the rotation angle of the related image relative to the reference image using a preset rotation angle algorithm, and performing rotation processing on the related image at the rotation angle.

[0092] This step is consistent with the above step S12 and will not be repeated here.

[0093] Step S43: performing a preset feature point extraction algorithm on the relevant image after the rotation process to obtain a corresponding relevant feature point set.

[0094] This step is consistent with the above step S13 and will not be repeated here.

[0095] Step S44: a preset number of matching processes of the related feature point sets and the reference feature point set are performed in parallel using a KD-Tree structure to obtain multiple groups of feature point pairs with the same name and the same features.

[0096] In the embodiment of the present invention, when the number of matching pairs between the relevant feature point set and the reference feature point set is relatively large, each matching pair can be regarded as an independent task, and a parallel KD-Tree structure can be used to improve the efficiency of image matching.

[0097] Step S45: determining the spatial position relationship between the corresponding related image and the reference image through the pair of feature points with the same name.

[0098] This step is consistent with the above step S15 and will not be repeated here.

[0099] Example 5

[0100] Figure 5 : is a flowchart of an oblique image matching method provided by Embodiment 5 of the present invention, the method comprising the following steps:

[0101] Step S51: using the reference image and the camera tilt angle, a preset screening algorithm is used to screen out a plurality of related images intersecting with the reference image from a plurality of tilted images.

[0102] This step is consistent with the above step S11 and will not be described again here.

[0103] Step S52: Calculating the rotation angle of the related image relative to the reference image by a preset rotation angle algorithm, and performing rotation processing of the related image at the rotation angle.

[0104] This step is consistent with the above step S12 and will not be repeated here.

[0105] Step S53: performing a preset feature point extraction algorithm on the relevant image after the rotation process to obtain a corresponding relevant feature point set.

[0106] This step is consistent with the above step S13 and will not be repeated here.

[0107] Step S54: matching the relevant feature point set with the reference feature point set of the reference image to obtain feature point pairs with the same name and the same features.

[0108] This step is consistent with the above step S14 and will not be repeated here.

[0109] Step S55: after obtaining the feature point pairs with the same name, using the RANSAC algorithm to eliminate mismatched point pairs in the feature point pairs with the same name.

[0110] In the embodiment of the present invention, when the intrinsic parameters of the camera are known, the RANSAC algorithm can also be used to remove the mismatched point pairs in the same-name feature point pairs. For example, the RANSAC algorithm can be used to calculate the basic matrix F between the reference image and the relevant image, which is a 3X3 matrix. l ,q r are a pair of matching points on two images respectively, then the F matrix satisfies: The 8-point method is used to calculate the F matrix parameters. Eight pairs of matching points are randomly selected from the matching point pairs for iterative calculation. The number of iterations is 2048, the internal point rate is 0.95, and the wrong matching point pairs are eliminated for the first time. The least squares algorithm is used to calculate the homography matrix H, q for the remaining matching points r =Hq l , remove the mismatched points that do not meet the threshold and get the final matching point pair.

[0111] Step S56: determining the spatial position relationship between the corresponding related image and the reference image through the pair of feature points with the same name.

[0112] This step is consistent with the above step S15 and will not be repeated here.

[0113] Figure 6 is a flow chart of another oblique image matching method provided by Embodiment 5 of the present invention, the method further comprising the following steps:

[0114] Step S57: Calculate the reprojection error of the feature point pairs with the same name, and use the reprojection error as a confidence evaluation of the spatial position relationship.

[0115] In the embodiment of the present invention, all pairs of feature points with the same name are used to establish a virtual camera for the two images and calculate the image position and posture. The intrinsic matrix E is calculated and used to obtain the translation T and rotation R. The reprojection error of the point is calculated using the rotation matrix and the translation matrix. When the camera matrix K is known:

[0116] P = K[R|T]; x = PX;

[0117] The mean error of the reprojection error of the matching point pairs is used as the above confidence evaluation.

[0118] Example 6

[0119] Figure 7 It is a structural schematic diagram of an inclined image matching device provided in Example 6 of the present invention.

[0120] The oblique image matching device 700 comprises:

[0121] The related image screening module 710 is used to screen out a plurality of related images intersecting with the reference image from a plurality of tilted images by using a reference image and a camera tilt angle through a preset screening algorithm;

[0122] A rotation angle calculation module 720, configured to calculate the rotation angle of the related image relative to the reference image by using a preset rotation angle algorithm, and perform rotation processing of the related image at the rotation angle;

[0123] A feature point extraction module 730 is used to process the relevant image after the rotation process using a preset feature point extraction algorithm to obtain a corresponding relevant feature point set;

[0124] A feature point matching module 740 is used to match the relevant feature point set with the reference feature point set of the reference image to obtain feature point pairs with the same name and the same features;

[0125] The position relationship determination module 750 is used to determine the spatial position relationship between the corresponding related image and the reference image through the pair of feature points with the same name.

[0126] In the embodiment of the present invention, a more detailed functional description of each of the above modules can be referred to the corresponding parts of the aforementioned embodiment, which will not be repeated here.

[0127] In addition, the present invention also provides a computer device, which includes a memory and a processor. The memory can be used to store a computer program, and the processor runs the computer program to enable the computer device to execute the functions of each module in the above method or the above oblique image matching device.

[0128] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the computer device (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0129] This embodiment also provides a computer storage medium for storing a computer program used in the above-mentioned computer device.

[0130] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and structure diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in an alternative implementation, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the structure diagram and / or the flow chart, and the combination of boxes in the structure diagram and / or the flow chart, can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0131] In addition, the functional modules or units in the various embodiments of the present invention may be integrated together to form an independent part, or each module may exist independently, or two or more modules may be integrated to form an independent part.

[0132] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a smart phone, a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0133] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for matching oblique images, characterized in that: include: Using the reference image and the camera tilt angle, a plurality of related images intersecting with the reference image are screened out from the plurality of tilted images through a preset screening algorithm; Calculating the rotation angle of the related image relative to the reference image by a preset rotation angle algorithm, and performing rotation processing of the related image by the rotation angle; Processing the rotated related images with a preset feature point extraction algorithm to obtain a corresponding related feature point set; Matching the relevant feature point set with the reference feature point set of the reference image to obtain feature point pairs with the same name and the same features; The method of using the reference image and the camera tilt angle to select a plurality of related images intersecting with the reference image from a plurality of tilted images through a preset screening algorithm comprises: Calculating the exterior orientation elements of the oblique image according to the camera tilt angle and the initial exterior orientation elements of the downward-looking image; Calculating the coverage of the oblique image according to the exterior orientation elements; Determine the oblique image whose coverage range intersects with the range of the reference image as the relevant image; The matching of the relevant feature point set with the reference feature point set of the reference image comprises: A KD-Tree structure is used to perform a matching process of a preset number of related feature point sets and a reference feature point set in parallel to obtain multiple groups of feature point pairs with the same name and same features; The spatial position relationship between the corresponding related image and the reference image is determined through the pair of feature points with the same name.

2. The oblique image matching method according to claim 1, characterized in that: The preset rotation angle algorithm is an ASIFT algorithm, and the step of calculating the rotation angle of the related image relative to the reference image by using the preset rotation angle algorithm includes: Performing downsampling processing on the reference image and the related images at a preset ratio; After downsampling, extracting initial matching feature points from the reference image and the related image using the ASIFT algorithm; Using the RANSAC algorithm to remove gross error points in the initial matching feature points to obtain matching feature points; The rotation angle of the related image relative to the reference image is calculated according to the matching feature points.

3. The oblique image matching method according to claim 1, characterized in that: The preset feature point extraction algorithm is the Hessian-Affine algorithm.

4. The oblique image matching method according to claim 1, characterized in that: Also includes: After obtaining the feature point pairs with the same name, the RANSAC algorithm is used to eliminate the mismatched point pairs in the feature point pairs with the same name.

5. The oblique image matching method according to claim 1, characterized in that: Also includes: The reprojection error of the pair of feature points with the same name is calculated, and the reprojection error is used as a confidence evaluation of the spatial position relationship.

6. An oblique image matching device, characterized in that: include: A related image screening module, used to screen out a plurality of related images intersecting with the reference image from a plurality of oblique images by using a reference image and a camera tilt angle through a preset screening algorithm; A rotation angle calculation module, used to calculate the rotation angle of the related image relative to the reference image by a preset rotation angle algorithm, and perform rotation processing of the related image at the rotation angle; A feature point extraction module is used to process the relevant image after the rotation processing with a preset feature point extraction algorithm to obtain a corresponding relevant feature point set; A feature point matching module is used to match the relevant feature point set with the reference feature point set of the reference image to obtain feature point pairs with the same name and the same features; The related image screening module is used to screen out multiple related images intersecting with the reference image from multiple oblique images by using the reference image and the camera tilt angle through a preset screening algorithm, including: Calculating the exterior orientation elements of the oblique image according to the camera tilt angle and the initial exterior orientation elements of the downward-looking image; Calculating the coverage of the oblique image according to the exterior orientation elements; Determine the oblique image whose coverage range intersects with the range of the reference image as the relevant image; The feature point matching module is used to match the relevant feature point set with the reference feature point set of the reference image, including: A KD-Tree structure is used to perform a matching process of a preset number of related feature point sets and a reference feature point set in parallel to obtain multiple groups of feature point pairs with the same name and same features; The position relationship determination module is used to determine the spatial position relationship between the corresponding related image and the reference image through the pair of feature points with the same name.

7. A computer device, characterized in that: The invention comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is run on the processor, the oblique image matching method according to any one of claims 1 to 5 is executed.

8. A readable storage medium, characterized in that: The device stores a computer program, which executes the oblique image matching method according to any one of claims 1 to 5 when running on a processor.

Images