Method, system, device and medium for expanding positioning map features

Abstract

The present application relates to a method, system, device and medium for expanding positioning map features, wherein the method includes: obtaining a source image of a positioning map, identifying expansion points of global features of the source image, and generating a list of patches and true value candidate frames; verifying the generated patches, and clustering the features of the patches that pass the verification to obtain available patch global features, which are expanded into the global features of the positioning map; performing a global query in the global features of the expanded positioning map based on the global features of the query image, obtaining candidate image labels, and completing the visual positioning of the image. Through this application, the problem of low accuracy of positioning using a purely visual hierarchical positioning algorithm is solved, and the success rate of positioning is effectively improved.

Description

Technical Field

[0001] The present application relates to the field of three-dimensional reconstruction technology, and in particular to a method, system, device and medium for expanding positioning map features. Background Art

[0002] When reconstructing three-dimensional maps of various scenes, available features in pure visual scene images are usually extracted to generate positioning maps for visual positioning of AR devices.

[0003] In related technologies, a localization map is a localizable map that includes visual features. Its main elements include: (1) accurate and complete three-dimensional geometric structure of the scene, where the geometric structure can be represented by a three-dimensional point cloud (3D points). (2) rich and highly recognizable visual features, which can be divided into global features and local features (2D feature points) according to the description level of the features. (3) complete and accurate 2D-3D correspondence.

[0004] After generating the scene positioning map, it is necessary to automatically visually position the device in the positioning map using a visual positioning algorithm. For visual positioning problems in large-scale scenarios, a hierarchical positioning method is generally used: that is, after receiving the query image, the global features and local features of the query image are extracted; then, based on the global features, several global features that are closest to it are quickly found in the positioning map, thereby finding the map image (candidates, also known as candidate frames) that is closest to the query image, and filtering out those less similar and irrelevant map images; finally, using the local features of the query image, one by one, they are matched with the local features of the candidate frames to find the most similar batch of 2D points, and then find the corresponding 3D points; when there are enough correspondences between the 2D points of the query image and the 3D points in the map, the 6DOF pose of the query image in the positioning map can be calculated.

[0005] As can be seen from the above, one of the key factors for successful positioning in purely visual hierarchical localization algorithms is the effectiveness of global queries. Local matching can only be successful if the accuracy of candidate frames found in the global query is high. However, in real-world applications, especially in open and complex environments like large shopping malls, directly using the global features of the query image and the map image for global search often produces suboptimal results and often leads to positioning failure.

[0006] Currently, no effective solution has been proposed to the problem of low accuracy in positioning using purely visual hierarchical positioning algorithms in related technologies. Summary of the Invention

[0007] The embodiments of the present application provide a method, system, device and medium for expanding positioning map features to at least solve the problem of low accuracy of positioning using a purely visual hierarchical positioning algorithm in the related art.

[0008] In a first aspect, an embodiment of the present application provides a method for expanding positioning map features, the method comprising:

[0009] Obtaining a source image of the positioning map, performing global feature expansion point recognition on the source image, and generating a patch and a true value candidate frame list;

[0010] Verify the generated patches, and perform feature clustering on the verified patches to obtain available patch global features, which are expanded into the global features of the positioning map;

[0011] Based on the global features of the query image, a global query is performed in the global features of the expanded positioning map to obtain candidate image labels.

[0012] In some embodiments, identifying expansion points of global features on the source image includes:

[0013] Screening the source images and selecting source images with ambiguous global features as expansion points;

[0014] Alternatively, the source images are not screened, and all source images are added to the expansion list as expansion points.

[0015] In some embodiments, screening the source images and selecting source images with global feature ambiguity as expansion points includes:

[0016] Obtaining a list of true value candidate frames, performing a global query in the positioning map according to the global features of the source image, and calculating the accuracy of the global query;

[0017] When the accuracy is lower than a preset threshold, the source image has global feature ambiguity and is a global feature point to be expanded.

[0018] In some embodiments, generating a truth candidate frame list includes:

[0019] Performing local feature matching and two-view geometry verification on the query image and all map images in the positioning map;

[0020] When the verification passes and the number of matches is greater than a preset threshold, the map images that have passed the verification and matching are sorted in descending order according to the number of matches of local features to obtain a true value candidate frame list of the query image.

[0021] In some embodiments, verifying the generated patch includes:

[0022] Performing a global feature query on the patch in the validation set, and mapping the feature query results to global candidate frame results based on the mapping relationship between the patch and the map image, wherein the validation set includes: global features of all patch images and global features of all source images;

[0023] Compare the global candidate frame result with the true value candidate frame list of the map image corresponding to the patch, and calculate the query accuracy;

[0024] If the accuracy rate exceeds the preset threshold, the patch is considered to have global discrimination and can be retained, otherwise it should be discarded.

[0025] In some embodiments, patches that have passed verification are clustered to obtain available global patch features including:

[0026] The patches that have passed the verification are classified by features through hierarchical clustering, and the patch images corresponding to the features in each category are checked to see if they are close enough. If not, the corresponding patch is deleted. If so, it is retained and the representative patch of each category is selected.

[0027] In some embodiments, performing a global query on the global features of the expanded positioning map based on the global features of the query image to obtain candidate image tags includes:

[0028] If the candidate image found is the original map image, no further operation is required;

[0029] If the candidate image found is a patch image, the original map image corresponding to the patch image needs to be replaced in the candidate image list.

[0030] In a second aspect, an embodiment of the present application provides a system for expanding positioning map features, the system comprising:

[0031] An expansion module is used to obtain a source image of a positioning map, identify expansion points of global features on the source image, and generate a patch and a true value candidate frame list;

[0032] A verification and clustering module is used to verify the generated patches and perform feature clustering on the patches that pass the verification to obtain available patch global features and expand them into the global features of the positioning map;

[0033] The query module is used to perform a global query in the global features of the expanded positioning map according to the global features of the query image to obtain candidate image labels.

[0034] In a third aspect, an embodiment of the present application provides an electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the method for expanding positioning map features as described in the first aspect above is implemented.

[0035] In a fourth aspect, an embodiment of the present application provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for expanding positioning map features as described in the first aspect above.

[0036] Compared with related technologies, the positioning map feature expansion method provided in the embodiments of the present application obtains the source image of the positioning map, identifies the expansion points of the global features of the source image, and generates a patch and a list of true value candidate frames; verifies the generated patch, and performs feature clustering on the patches that pass the verification to obtain available patch global features, which are expanded into the global features of the positioning map; performs a global query in the global features of the expanded positioning map based on the global features of the query image, obtains candidate image labels, and completes the visual positioning of the image.

[0037] This application proposes a global feature expansion scheme based on image patches to achieve visual positioning and viewing of map images. Specifically, for a map image or an image from another source, through the solution of this application, several sub-images (patches) of different sizes and positions are extracted from the original image, and the most globally discriminative effective patch is selected from them, and its global features are expanded to the global features of the positioning map. The global features of the expanded positioning map include not only the original global feature set, but also the global features of the effective patch, thereby providing more discriminative query perspectives for the positioning map, greatly improving the accuracy of global search, solving the problem of low accuracy of positioning through purely visual hierarchical positioning algorithms, and effectively improving the success rate of positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0039] Figure 1 is a flowchart of a method for expanding positioning map features according to an embodiment of the present application;

[0040] Figure 2is a structural block diagram of a system for expanding positioning map features according to an embodiment of the present application;

[0041] Figure 3 Schematic diagram of the internal structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present application. In addition, it can also be understood that although the efforts made in this development process may be complex and lengthy, for ordinary technicians in the field related to the contents disclosed in the present application, some changes such as design, manufacturing or production based on the technical contents disclosed in the present application are only conventional technical means and should not be understood as the contents disclosed in the present application being insufficient.

[0043] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments unless there is a conflict.

[0044] Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by a person of ordinary skill in the technical field to which this application belongs. The words "one", "a", "the" and the like used in this application do not indicate a limit on quantity and may indicate the singular or plural. The terms "include", "comprise", "have" and any variations thereof used in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units that are inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The word "multiple" used in this application means greater than or equal to two. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The terms "first", "second", "third" and the like involved in this application are merely used to distinguish similar objects and do not represent a specific ordering of the objects.

[0045] This embodiment provides a method for expanding positioning map features. Figure 1 is a flow chart of a method for expanding positioning map features according to an embodiment of the present application. Figure 1 As shown, the process includes the following steps:

[0046] Step S101: Acquire a source image of a positioning map, perform global feature expansion point recognition on the source image, and generate a patch and a true value candidate frame list. It should be noted that, in this embodiment, generating a patch includes: generating a patch image, and generating global features of the patch image.

[0047] Acquire the source image of the localization map and use it as the scene image for subsequent patch generation and global feature expansion. The source images acquired in this implementation include internal source images and external source images. Internal source images are images used during map construction, i.e., map images; external source images are other images taken separately, i.e., images that have not been used in map reconstruction.

[0048] Furthermore, the global feature expansion point recognition is performed on the source image obtained above, and a patch and a true value candidate frame list are generated, as follows:

[0049] For some large scene maps, since global feature ambiguity does not exist everywhere, in order to save the time of the expansion process, the acquired source images can be screened, and source images with global feature ambiguity can be selected as expansion points. Patches and corresponding true value candidate frame lists are generated only for these source images; alternatively, the above-mentioned source image screening can be omitted, and all source images can be added to the expansion list as expansion points to generate patches and corresponding true value candidate frame lists;

[0050] Among them, the specific steps of screening the source images and selecting the source images with global feature ambiguity as expansion points include: obtaining a list of true value candidate frames, performing a global query in the positioning map based on the global features of the source images, and calculating the accuracy of the global query, that is, calculating the probability that the candidate frames obtained by the global query are in the list of true value candidate frames; when the accuracy is lower than a preset threshold, it can be considered that the source image has global feature ambiguity, that is, the scene location captured by the image has global feature ambiguity, and it is a global feature point to be expanded.

[0051] Preferably, in this embodiment, the method for generating a list of ground-truth candidate frames includes: performing local feature matching and two-view geometry verification on the query image and all map images in the positioning map; when the verification passes and the number of matches exceeds a preset threshold, the matched image is considered similar to the query image. At this point, the map images that have passed the verification and matching are sorted in descending order according to the number of local feature matches to obtain a list of ground-truth candidate frames for the query image. This list of ground-truth candidate frames can be used to evaluate the global query results, i.e., calculate the accuracy of the global query.

[0052] Assume that the query image is imgQ, the query result is resQ = {img1, img2, ..., imgX}, the length of the query result is K, and the true value candidate frame list is resGT = {img1, img2, ..., imgY}. The calculation formula for the accuracy of the global query result is shown in the following formula (1):

[0053]

[0054] where |{·}| represents the number of elements in the set.

[0055] Preferably, there are many methods for generating patches in this embodiment. One of the more common methods is to use patch templates of different sizes and move them in the source image with a certain step size. Each time a patch moves to a new position, the corresponding image portion is deducted as a patch image. In addition, each patch image can independently extract global features using a global feature extractor, such as extracting NetVlad features. It should be noted that each patch will record the original image in its corresponding positioning map.

[0056] Step S102: Verify the generated patches and perform feature clustering on the patches that pass the verification to obtain available patch global features, which are then expanded into the global features of the positioning map.

[0057] The number of patches generated by the above step S101 is large. For example, the original image of each positioning map can generate dozens or even hundreds of patches. Among these patches, there is a lot of redundancy. Some patches even have high global ambiguity. If they are expanded to the positioning map, they will confuse the results of the global query. Therefore, in this embodiment, it is necessary to verify the generated patches, select those with global discrimination (i.e., low global ambiguity), and filter out those with high global ambiguity. Preferably, the verification of the generated patches includes:

[0058] First, for each patch, perform a global feature query on the validation set, and map the feature query results to global candidate frame results based on the mapping relationship between the patch and the map image. The validation set includes: the global features of all patch images and the global features of all source images;

[0059] Next, the global candidate frame results are compared with the true candidate frame list of the map image corresponding to the patch, and the query accuracy is calculated;

[0060] If the accuracy rate exceeds a preset threshold, such as 50%, the patch is considered to have global discrimination and can be retained; otherwise, it should be discarded.

[0061] Furthermore, the patches that have passed the verification are clustered to obtain available patch global features, which are then expanded into the global features of the positioning map. Preferably, the patches that have passed the verification are clustered to obtain available patch global features, which include:

[0062] (1) Gather all the global features of the patches that have passed the verification, and use hierarchical clustering or other clustering methods to divide the global features of the patches into multiple categories. The features within each category are relatively close.

[0063] (2) Check whether the patch images corresponding to the global features of the patches in each category are close enough. If not, the corresponding patches are deleted. If so, they are retained. The principles for judging whether the patch images are close enough include: whether the two patches correspond to the same map image, or whether the distance between the camera positions of the map images corresponding to the two patches is less than a preset threshold;

[0064] (3) For each remaining category, one patch can be selected as the representative of the category and expanded into the global features of the positioning map. The method for selecting the representative patch includes: selecting the patch with the highest accuracy of the global query result as the representative of the category.

[0065] It should be noted that after the global features of the positioning map are expanded through the above process, the corresponding original map images of the patches generated from them need to be recorded for both internal and external source images. For internal source images, the reverse recording can be performed when the patch is generated. That is, the map image corresponding to the patch generated from the internal source image is the source image itself. For external source images, the map image corresponding to the patch generated from it should be the ground truth candidate frame image that ranks first in the map image of the external source image.

[0066] This embodiment selects the most globally discriminative patch from the source image and expands its global features into the global features of the positioning map. This provides the positioning map with more discriminative queryable perspectives, greatly improving the accuracy of global search and, in turn, the success rate of positioning.

[0067] Step S103 : performing a global query in the global features of the expanded positioning map based on the global features of the query image to obtain candidate image labels.

[0068] Preferably, after the global features of the positioning map are expanded through the above step S102, the positioning resource will include two parts of global features: the global features of the original map image and the global features of the patch image. When a global query is performed in the global features of the expanded positioning map based on the global features of the query image, there are two possible candidate image labels directly obtained: if the candidate image queried is the original map image, no further operation is required; if the candidate image queried is the patch image, the original map image corresponding to the patch image should be replaced in the candidate image list; if an original map image has already been added to the list, the search should continue for the next candidate.

[0069] Furthermore, if the expanded patch global features are confirmed to have high global discriminability, we can choose to use only or prioritize the patch global features for global queries, while discarding or partially discarding the global features of the original map image. This can be beneficial in some cases, such as when the global features of the original map image are highly ambiguous, thus improving the accuracy of global searches.

[0070] Through steps S101 to S103 above, this embodiment extracts several sub-images (patches) of different sizes and positions from the original image, selects the most globally discriminative valid patch, and expands its global features into the global features of the positioning map. The expanded global features of the positioning map include not only the original global feature set but also the global features of the valid patches, thereby providing the positioning map with more discriminative query perspectives, greatly improving the accuracy of global search, and solving the problem of low accuracy in positioning using purely visual hierarchical positioning algorithms, effectively improving the success rate of positioning.

[0071] It should be noted that the steps shown in the above process or the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0072] This embodiment also provides a system for expanding positioning map features, which is used to implement the above-mentioned embodiments and preferred embodiments. Details already described will not be repeated here. As used below, the terms "module," "unit," "subunit," etc. may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0073] Figure 2 This is a structural block diagram of a system for expanding positioning map features according to an embodiment of the present application. Figure 2 As shown, the system includes an expansion module 21, a verification clustering module 22 and a query module 23:

[0074] The expansion module 21 is used to obtain the source image of the positioning map, identify the expansion points of the global features of the source image, and generate a patch and a list of true value candidate frames; the verification and clustering module 22 is used to verify the generated patches and perform feature clustering on the patches that pass the verification to obtain usable patch global features and expand them into the global features of the positioning map; the query module 23 is used to perform a global query in the global features of the expanded positioning map based on the global features of the query image to obtain candidate image labels.

[0075] Through the above system, this embodiment extracts several sub-images (patches) of different sizes and positions from the original image, selects the most globally discriminative valid patch, and expands its global features into the global features of the positioning map. The expanded global features of the positioning map include not only the original global feature set but also the global features of the valid patches, thereby providing the positioning map with more discriminative query perspectives, greatly improving the accuracy of global searches, and solving the low accuracy problem of positioning using purely visual hierarchical positioning algorithms, effectively improving the success rate of positioning.

[0076] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be repeated here.

[0077] Furthermore, it should be noted that each of the aforementioned modules can be either a functional module or a program module, and can be implemented via software or hardware. For modules implemented via hardware, each of the aforementioned modules can be located in the same processor; or each of the aforementioned modules can be located in different processors in any combination.

[0078] This embodiment further provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0079] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0080] In addition, in conjunction with the method for expanding positioning map features in the above embodiments, embodiments of the present application may provide a storage medium for implementation. The storage medium stores a computer program; when the computer program is executed by a processor, it implements any of the methods for expanding positioning map features in the above embodiments.

[0081] In one embodiment, a computer device is provided, which may be a terminal. The computer device includes a processor, memory, a network interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for expanding positioning map features is implemented. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen. The input device of the computer device may be a touch layer covering the display screen, or may be a key, trackball, or touchpad provided on the computer device housing, or may be an external keyboard, touchpad, or mouse.

[0082] In one embodiment, Figure 3 is a schematic diagram of the internal structure of an electronic device according to an embodiment of the present application, such as Figure 3 As shown, an electronic device is provided, which may be a server, and its internal structure diagram may be as shown in FIG. Figure 3 As shown. The electronic device includes a processor, a network interface, an internal memory, and a non-volatile memory connected via an internal bus, wherein the non-volatile memory stores an operating system, a computer program, and a database. The processor is used to provide computing and control capabilities, the network interface is used to communicate with external terminals via a network connection, the internal memory is used to provide an environment for the operation of the operating system and the computer program, and when executed by the processor, the computer program implements a method for expanding positioning map features. The database is used to store data.

[0083] Those skilled in the art will understand that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the electronic device to which the solution of the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0084] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, which can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0085] Those skilled in the art should understand that the various technical features of the above-described embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the various technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0086] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for expanding positioning map features, characterized in that: The method comprises: Obtaining a source image of the positioning map, performing global feature expansion point recognition on the source image, and generating a patch and a true value candidate frame list; The generated patches are verified, and the patches that pass the verification are clustered to obtain available patch global features, which are expanded into the global features of the positioning map. The available patch global features obtained by clustering the patches that pass the verification include: Perform feature classification on the patches that have passed the verification through hierarchical clustering, and check whether the patch images corresponding to the features in each category are close enough. If not, the corresponding patch is deleted. If so, it is retained and the representative patch of each category is selected; Based on the global features of the query image, a global query is performed in the global features of the expanded positioning map to obtain candidate image labels.

2. The method according to claim 1, characterized in that Performing global feature expansion point recognition on the source image includes: Screening the source images and selecting source images with ambiguous global features as expansion points; Alternatively, the source images are not screened, and all source images are added to the expansion list as expansion points.

3. The method according to claim 2, characterized in that Screening the source images and selecting source images with global feature ambiguity as expansion points includes: Obtaining a list of true value candidate frames, performing a global query in the positioning map according to the global features of the source image, and calculating the accuracy of the global query; When the accuracy is lower than a preset threshold, the source image has global feature ambiguity and is a global feature point to be expanded.

4. The method according to claim 1, wherein Generating a true value candidate frame list includes: Performing local feature matching and two-view geometry verification on the query image and all map images in the positioning map; When the verification passes and the number of matches is greater than a preset threshold, the map images that have passed the verification and matching are sorted in descending order according to the number of matches of local features to obtain a true value candidate frame list of the query image.

5. The method according to claim 1, wherein Verifying the generated patch includes: Performing a global feature query on the patch in the validation set, and mapping the feature query results to global candidate frame results based on the mapping relationship between the patch and the map image, wherein the validation set includes: global features of all patch images and global features of all source images; Compare the global candidate frame result with the true value candidate frame list of the map image corresponding to the patch, and calculate the query accuracy; If the accuracy rate exceeds the preset threshold, the patch is considered to have global discrimination and can be retained, otherwise it should be discarded.

6. The method according to claim 1, characterized in that Based on the global features of the query image, a global query is performed in the global features of the expanded positioning map, and the candidate image labels obtained include: If the candidate image found is the original map image, no further operation is required; If the candidate image found is a patch image, the original map image corresponding to the patch image needs to be replaced in the candidate image list.

7. A system for expanding positioning map features, characterized in that: The system comprises: An expansion module is used to obtain a source image of a positioning map, identify expansion points of global features on the source image, and generate a patch and a true value candidate frame list; The verification clustering module is used to verify the generated patches and perform feature clustering on the patches that pass the verification to obtain available patch global features, which are then expanded into the global features of the positioning map. The features of the patches that pass the verification are clustered to obtain available patch global features, including: Perform feature classification on the patches that have passed the verification through hierarchical clustering, and check whether the patch images corresponding to the features in each category are close enough. If not, the corresponding patch is deleted. If so, it is retained and the representative patch of each category is selected; The query module is used to perform a global query in the global features of the expanded positioning map according to the global features of the query image to obtain candidate image labels.

8. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to run the computer program to perform the method for expanding positioning map features according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium stores a computer program, wherein: The computer program is configured to execute the method for expanding positioning map features according to any one of claims 1 to 6 when running.

Images