POI (Point of Interest) duplication judgment method, electronic equipment, storage medium and program product

Abstract

The embodiment of the invention provides an interest point duplication judgment method, electronic equipment, a storage medium and a program product. The method relates to the field of artificial intelligence and the field of maps, and comprises the following steps: in response to a repetition judgment instruction for a to-be-processed interest point, selecting candidate interest points similar to the to-be-processed interest point from map interest points; intercepting a first interested image and a second interested image in the map by respectively taking the to-be-processed interested point and the candidate interested point as centers; the first text information of the to-be-processed interest point, the first interested image, the second text information of the candidate interest point and the second interested image are subjected to duplication judgment processing through the multi-modal network model, and a duplication judgment result is obtained and used for representing whether the to-be-processed interest point and the candidate interest point are repeated or not. The method is used for achieving the effect of improving the accuracy of interest point duplicate judgment.

Description

Technical Field

[0001] This application relates to the fields of artificial intelligence and maps, and in particular, to a method for judging duplicate points of interest, an electronic device, a storage medium, and a program product. Background Art

[0002] Judging duplicate points of interest on a map can reduce data redundancy, improve search efficiency, and save storage space, which is an important part of map data processing.

[0003] In related technologies, duplicate points are judged by the similarity of the names of points of interest. However, there are points of interest with similar names but different natures, resulting in low accuracy of duplicate point judgment. Summary of the Invention

[0004] The method for judging duplicate points of interest, electronic device, storage medium, and program product provided by the embodiments of this application are used to achieve the effect of improving the accuracy of duplicate point judgment.

[0005] In a first aspect, an embodiment of this application provides a method for judging duplicate points of interest, including:

[0006] In response to a duplicate point judgment instruction for a to-be-processed point of interest, select candidate points of interest similar to the to-be-processed point of interest from map points of interest;

[0007] Taking the to-be-processed point of interest and the candidate points of interest as centers respectively, intercept a first region of interest image and a second region of interest image in the map;

[0008] Perform duplicate point judgment processing on the first text information, the first region of interest image of the to-be-processed point of interest, the second text information, and the second region of interest image of the candidate points of interest through a multi-modal network model to obtain a duplicate point judgment result, where the duplicate point judgment result is used to indicate whether the to-be-processed point of interest and the candidate points of interest are duplicates.

[0009] In a second aspect, an embodiment of this application provides a device for judging duplicate points of interest, including:

[0010] A selection module, configured to select candidate points of interest similar to the to-be-processed point of interest from map points of interest in response to a duplicate point judgment instruction for the to-be-processed point of interest;

[0011] An image interception module, configured to intercept a first region of interest image and a second region of interest image in the map with the to-be-processed point of interest and the candidate points of interest as centers respectively;

[0012] A multi-modal processing module, configured to perform duplicate point judgment processing on the first text information, the first region of interest image of the to-be-processed point of interest, the second text information, and the second region of interest image of the candidate points of interest through a multi-modal network model to obtain a duplicate point judgment result, where the duplicate point judgment result is used to indicate whether the to-be-processed point of interest and the candidate points of interest are duplicates.

[0013] In a third aspect, an embodiment of the present application provides an electronic device, including: a memory and a processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementation manners of the first aspect.

[0014] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the computer-executable instructions are executed by a processor, they are used to implement the above first aspect and / or various possible implementation manners of the first aspect.

[0015] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the above first aspect and / or various possible implementation manners of the first aspect.

[0016] The method, electronic device, storage medium and program product for determining duplicate points of interest provided by the embodiments of the present application screen candidate points of interest similar to the point of interest to be processed in the map of points of interest, intercept a first image of interest centered on the point of interest to be processed and a second image of interest centered on the candidate point of interest in the map, introduce high-order visual features around the point of interest to be processed and the candidate point of interest, and perform duplicate point determination processing on the first text information, the first image of interest of the point of interest to be processed, the second text information of the candidate point of interest and the second image of interest through a multimodal network model to obtain a duplicate point determination result. By fusing the text and image data of the point of interest to be processed and the candidate point of interest, the richness of the data used for duplicate point determination of points of interest is improved, and the accuracy of duplicate point determination of points of interest is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0018] Figure 1 It is a schematic diagram of the scenario of the method for determining duplicate points of interest provided by the present application;

[0019] Figure 2 It is a schematic flow chart of the method for determining duplicate points of interest provided by the present application Figure 1 ;

[0020] Figure 3 It is a schematic diagram of the intersection situation between the point of interest and the area of interest provided by the present application Figure 1 ;

[0021] Figure 4 It is a schematic diagram of the intersection situation between the point of interest and the area of interest provided by the present application Figure 2 ;

[0022] Figure 5 Schematic diagram of the intersection between the point of interest and the surface of interest provided by this application Figure 3 ;

[0023] Figure 6 Schematic diagram of the intersection between the point of interest and the surface of interest provided by this application Figure 4 ;

[0024] Figure 7 Schematic diagram of the intersection between the point of interest and the surface of interest provided by this application Figure 5 ;

[0025] Figure 8 Schematic diagram for determining the interesting image corresponding to the point of interest provided by this application;

[0026] Figure 9 Flow schematic of the method for judging duplicate points of interest provided by this application Figure 2 ;

[0027] Figure 10 Flow schematic diagram of the device for judging duplicate points of interest provided by this application;

[0028] Figure 11 Schematic diagram of the structure of the electronic device provided by this application.

[0029] Through the above-mentioned drawings, specific embodiments of this application have been shown, and there will be more detailed descriptions hereinafter. These drawings and textual descriptions are not intended to limit the scope of the concept of this application in any way, but to illustrate the concept of this application to those skilled in the art by referring to specific embodiments. Detailed Description of the Embodiments

[0030] Here, the exemplary embodiments will be described in detail, and the examples are shown in the drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. On the contrary, they are merely examples of devices and methods consistent with some aspects of this application as detailed in the appended claims.

[0031] The method for judging duplicate points of interest provided by the embodiments of this application can be applied to the application environment as Figure 1 shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or placed in the cloud or other network servers.

[0032] Among them, the terminal 102 can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, Internet of Things devices, and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0033] The following uses specific embodiments to elaborate in detail on the technical solution of this application and how the technical solution of this application solves the above technical problems. These several specific embodiments below can be combined with each other, and for the same or similar concepts or processes, they may not be repeated in some embodiments. The following will describe the embodiments of this application in conjunction with the accompanying drawings.

[0034] Figure 2 Flow schematic of the point of interest duplication determination method provided by this application Figure 1 , the point of interest duplication determination method can be applied to an electronic device, and the electronic device can be Figure 1 the terminal or server in Figure 2 As shown in

[0035] S201. In response to a duplication determination instruction for a to-be-processed point of interest, select candidate points of interest similar to the to-be-processed point of interest from the map points of interest.

[0036] Among them, a point of interest (POI, Point of Interest) is a marker for a specific geographical location, such as a restaurant, museum, store, school, etc.

[0037] Map points of interest refer to all points of interest in the map, and point of interest duplication determination is to determine whether the to-be-processed point of interest duplicates with the map points of interest.

[0038] Optionally, when updating the map, use the newly added points of interest as the to-be-processed points of interest, and use the original points of interest in the map as the map points of interest; optionally, when regularly checking the map data, use the original points of interest in the map as the to-be-processed points of interest in turn, and use the other points of interest in the original points of interest except the to-be-processed points of interest as the map points of interest.

[0039] Among them, candidate points of interest are points of interest similar to the to-be-processed point of interest in the map points of interest, and the number of candidate points of interest can be multiple.

[0040] Specifically, the user issues a duplicate-check instruction for the interest point to be processed. There are various ways to issue a duplicate-check instruction for the interest point to be processed. For example, the map service platform has a duplicate-check function. The interest point to be processed is selected through the duplicate-check function, and the duplicate-check instruction is triggered by the duplicate-check button corresponding to the duplicate-check function. The embodiments of the present application do not limit the way of issuing the duplicate-check instruction.

[0041] In response to the duplicate-check instruction, the electronic device determines the map interest points, obtains the first text information of the interest point to be processed, obtains the initial text information of the map interest points, and based on the first text information and the initial text information, selects candidate interest points similar to the interest point to be processed from the map interest points.

[0042] In a possible way, based on the first text information and the initial text information, selecting candidate interest points similar to the interest point to be processed from the map interest points includes: through a search engine, recalling the initial text information according to the first text information to obtain candidate interest points similar to the interest point to be processed.

[0043] In another possible way, based on the first text information and the initial text information, selecting candidate interest points similar to the interest point to be processed from the map interest points includes: through a search engine, recalling the initial text information according to the first text information to obtain the initial interest points, obtaining the interest surface where the initial interest points are located and the interest surface where the interest point to be processed is located. In the case where the interest surface where the initial interest points are located intersects with the interest surface where the interest point to be processed is located, the initial interest points are used as candidate interest points similar to the interest point to be processed.

[0044] S202: Respectively centered on the interest point to be processed and the candidate interest points, intercept the first interested image and the second interested image in the map.

[0045] Among them, the interested image is an image corresponding to the region of interest (ROI, Region of Interest). The region of interest is a geographical region within a certain radius centered on the interest point. The region of interest contains relevant information or features of the interest point, such as other locations, facilities or landscapes related to the interest point.

[0046] The first interested image is an interested image of the region of interest centered on the interest point to be processed; the second interested image is an interested image of the region of interest centered on the candidate interest points.

[0047] Specifically, the electronic device determines the interested radius of the interest point to be processed, and centered on the interest point to be processed, intercepts the first interested image in the map according to its interested radius; determines the interested radius of the candidate interest points, and centered on the candidate interest points, intercepts the second interested image in the map according to its interested radius.

[0048] Optionally, the radius of interest of the interest point to be processed can be determined based on the relevant information near the interest point to be processed in the map; the radius of interest of the candidate interest point can be determined based on the relevant information near the candidate interest point in the map.

[0049] Optionally, the radius of interest of the interest point to be processed and the radius of interest of the candidate interest point can be preset radii.

[0050] S203. Perform duplicate checking on the first text information, the first interesting image of the interest point to be processed, the second text information, and the second interesting image of the candidate interest point through a multimodal network model to obtain a duplicate checking result, where the duplicate checking result is used to indicate whether the interest point to be processed and the candidate interest point are duplicates.

[0051] Among them, the first text information is the text - type information of the interest point to be processed, and the first text information is used to describe various aspects of the interest point to be processed, including the name, address, category, etc. of the interest point to be processed; the second text information is the text - type information of the candidate interest point, and the second text information is used to describe various aspects of the candidate interest point, including the name, address, category, etc. of the candidate interest point; in practical applications, by comparing the first text information and the second text information, the differences between the interest point to be processed and the candidate interest point in terms of text - type information can be determined.

[0052] Multimodal refers to multiple types of data modalities. In the embodiments of the present application, text information and interesting images are multimodal data; the multimodal network model is used to process text information and interesting images to obtain a duplicate checking result.

[0053] The duplicate checking result can be that the interest point to be processed and the candidate interest point are duplicates, or the interest point to be processed and the candidate interest point are not duplicates.

[0054] Specifically, through the multimodal network model, extract the features of the first text information, the features of the first interesting image, the features of the second text information, and the features of the second interesting image, fuse the features of the first text information and the features of the first interesting image to obtain the fused features of the interest point to be processed, fuse the features of the second text information and the features of the second interesting image to obtain the fused features of the candidate interest point, and perform duplicate checking classification on the fused features of the interest point to be processed and the fused features of the candidate interest point to obtain the duplicate checking result.

[0055] The method for judging duplicate points of interest provided by the embodiments of the present application screens candidate points of interest similar to the to-be-processed point of interest from the map points of interest, intercepts a first image of interest centered on the to-be-processed point of interest and a second image of interest centered on the candidate point of interest in the map, introduces high-order visual features around the to-be-processed point of interest and the candidate point of interest, and performs duplicate point judgment processing on the first text information, the first image of interest of the to-be-processed point of interest, the second text information, and the second image of interest of the candidate point of interest through a multimodal network model to obtain a duplicate point judgment result. By fusing the text and image data of the to-be-processed point of interest and the candidate point of interest, the richness of the data used for duplicate point judgment of points of interest is improved, and the accuracy of duplicate point judgment of points of interest is improved.

[0056] In some embodiments, selecting candidate points of interest similar to the to-be-processed point of interest from the map points of interest includes: performing a recall process on the map points of interest according to the to-be-processed point of interest to screen initial points of interest in the map points of interest; obtaining a first area of interest where the to-be-processed point of interest is located and a second area of interest where the initial point of interest is located; determining an intersection result according to a first positional relationship between the to-be-processed point of interest and the second area of interest, and a second positional relationship between the initial point of interest and the first area of interest; when the intersection result indicates that the to-be-processed point of interest and the initial point of interest belong to the intersection area of the first area of interest and the second area of interest, taking the initial point of interest as a candidate point of interest similar to the to-be-processed point of interest.

[0057] Among them, the initial point of interest is a part of the map points of interest; the area of interest (AOI) is a geographical area, which may include multiple points of interest; the first area of interest contains the to-be-processed point of interest, and the second area of interest contains the initial point of interest.

[0058] Among them, the first positional relationship can indicate whether the to-be-processed point of interest is within the second area of interest, and the second positional relationship can indicate whether the initial point of interest is within the first area of interest.

[0059] Specifically, the electronic device obtains the first text information of the to-be-processed point of interest, obtains the original text information of the map points of interest, and performs a rough recall on the map points of interest through a search engine according to the first text information and the original text information to obtain initial points of interest similar to the to-be-processed point of interest in the text dimension; in practical applications, the search engine can be an ES (Elastic Search) engine, and ES is a distributed, highly scalable, and highly real-time search and data analysis engine.

[0060] The electronic device can obtain the geographical location of the interest point to be processed (such as the GPS location), call the geographical information system interface, and use this geographical location as a query parameter to query the unique identifier of the interest point to be processed. Based on the mapping relationship between the unique identifier and the interest surface, the electronic device can obtain the first interest surface where the interest point to be processed is located according to the unique identifier of the interest point to be processed. Similarly, the electronic device can obtain the geographical location of the initial interest point, use the geographical location of the initial interest point as a query parameter, query the unique identifier of the initial interest point by calling the geographical information system interface, and then obtain the second interest surface where the initial interest point is located according to the mapping relationship between the unique identifier and the interest surface.

[0061] The electronic device determines the first positional relationship based on the interest point to be processed and the second interest surface, and determines the second positional relationship based on the initial interest point and the first interest surface. It should be noted that since the first interest surface contains the interest point to be processed and the second interest surface contains the initial interest point, when the first positional relationship indicates that the interest point to be processed is within the second interest surface and the second positional relationship indicates that the initial interest point is within the first interest surface, it means that there is an intersection area between the first interest surface and the second interest surface, and both the interest point to be processed and the initial interest point are within the intersection area.

[0062] In a possible implementation, the set of position points corresponding to the second interest surface is obtained, and it is determined whether the geographical location of the interest point to be processed belongs to the set of position points corresponding to the second interest surface. If it belongs, the first positional relationship is determined as the interest point to be processed being inside the second interest surface. If it does not belong, the first positional relationship is determined as the interest point to be processed not being inside the second interest surface.

[0063] The set of position points corresponding to the first interest surface is obtained, and it is determined whether the geographical location of the initial interest point belongs to the set of position points corresponding to the first interest surface. If it belongs, the second positional relationship is determined as the initial interest point being inside the first interest surface. If it does not belong, the second positional relationship is determined as the initial interest point not being inside the first interest surface.

[0064] In another possible implementation, after obtaining the first interest surface where the interest point to be processed is located and the second interest surface where the initial interest point is located, it further includes: taking the interest point to be processed as the origin, determining a first ray in any direction; obtaining a first boundary line that intersects with the first ray among the boundary lines of the second interest surface; determining a first winding number based on the first ray and the first boundary line; when the first winding number belongs to a preset winding number interval, determining the first positional relationship between the interest point to be processed and the second interest surface as the interest point to be processed being inside the second interest surface.

[0065] Among them, the boundary line of the second interest surface is the side of the polygon corresponding to the second interest surface; the preset winding number interval is a positive integer interval that does not include 0.

[0066] Specifically, taking the point of interest to be processed as the origin, a first ray in an arbitrary direction is determined. The second interest surface is regarded as a polygon, and the sides of the polygon corresponding to the second interest surface are used as the boundary lines of the second interest surface. It is judged one by one whether the extension line of the boundary line of the second interest surface intersects the first ray. If the extension line of a certain boundary line of the second interest surface intersects the first ray, then this boundary line is taken as the first boundary line. For each selected first boundary line, the starting point and the ending point of the first boundary line are obtained, a first vector is determined according to the starting point and the ending point, and a first detection vector is determined according to the point of interest to be processed and the starting point of the first boundary line. The winding count corresponding to the first boundary line is determined according to the first vector and the first detection vector, and the first winding number is determined according to the winding count corresponding to each first boundary line. If the first winding number is not 0 (belongs to the preset winding number interval), it is determined that the first positional relationship is that the point of interest to be processed is inside the second interest surface. If the first winding number is 0 (does not belong to the preset winding number interval), it is determined that the first positional relationship is that the point of interest to be processed is not inside the second interest surface.

[0067] Among them, both the starting point and the ending point of the first boundary line are the vertices of the polygon corresponding to the second interest surface. The first vector corresponds to the direction vector from the starting point to the ending point, and the first detection vector corresponds to the direction vector from the point of interest to be processed to the starting point. Calculate the cross product between the first vector and the first detection vector. If the cross product is positive, it means that the first vector rotates counterclockwise relative to the first detection vector, and the winding count is incremented by 1. If the cross product is negative, it means that the first vector rotates clockwise relative to the first detection vector, and the winding count is decremented by 1. The sum value of the winding counts respectively determined based on each first boundary line is statistically obtained to get the first winding number.

[0068] In the above embodiment, by using the first ray and the boundary line of the second interest surface, the first winding number is determined, and then the first positional relationship between the point of interest to be processed and the second interest surface can be quickly determined, improving the efficiency of point of interest duplicate judgment.

[0069] In another possible implementation, after obtaining the first interest surface where the point of interest to be processed is located and the second interest surface where the initial interest point is located, it further includes: taking the initial interest point as the origin, determining a second ray in an arbitrary direction; obtaining a second boundary line that intersects the second ray among the boundary lines of the first interest surface; determining a second winding number based on the second ray and the second boundary line; when the second winding number belongs to the preset winding number interval, determining that the second positional relationship between the initial interest point and the first interest surface is that the initial interest point is inside the first interest surface.

[0070] Among them, the boundary line of the first interest surface is the side of the polygon corresponding to the first interest surface; the preset winding number interval is a positive integer interval that does not include 0.

[0071] Specifically, taking the initial point of interest as the origin, a second ray in an arbitrary direction is determined. The first interest surface is regarded as a polygon, and the sides of the polygon corresponding to the first interest surface are used as the boundary lines of the first interest surface. It is determined one by one whether the extension line of the boundary line of the first interest surface intersects the second ray. If the extension line of a certain boundary line of the first interest surface intersects the second ray, then this boundary line is used as the second boundary line. For each selected second boundary line, the starting point and the ending point of the second boundary line are obtained, a second vector is determined according to the starting point and the ending point, and a second detection vector is determined according to the initial point of interest and the starting point of the second boundary line. The winding count corresponding to the second boundary line is determined according to the second vector and the second detection vector, and the second winding number is determined according to the winding count corresponding to each second boundary line. If the second winding number is not 0 (belonging to a preset winding number interval), it is determined that the second positional relationship is that the initial point of interest is inside the first interest surface. If the second winding number is 0 (not belonging to the preset winding number interval), it is determined that the second positional relationship is that the point of interest to be processed is not inside the first interest surface.

[0072] Wherein, both the starting point and the ending point of the second boundary line are vertices of the polygon corresponding to the first interest surface. The second vector corresponds to the direction vector from the starting point to the ending point, and the second detection vector corresponds to the direction vector from the initial point of interest to the starting point. The cross product between the second vector and the second detection vector is calculated. If the cross product is positive, it means that the second vector rotates counterclockwise relative to the second detection vector, and the winding count is incremented by 1. If the cross product is negative, it means that the second vector rotates clockwise relative to the second detection vector, and the winding count is decremented by 1. The sum value of the winding counts respectively determined based on each second boundary line is statistically obtained to get the second winding number.

[0073] In practical applications, the solution for determining whether a point of interest is inside an interest surface by using the winding number is also known as the WindingNumber method, that is, the winding number method.

[0074] In the above embodiment, by using the second ray and the boundary line of the first interest surface, the second winding number is determined, and then the second positional relationship between the initial point of interest and the first interest surface can be quickly determined, improving the efficiency of point-of-interest duplicate determination.

[0075] In some embodiments, according to the first positional relationship between the point of interest to be processed and the second interest surface, and the second positional relationship between the initial point of interest and the first interest surface, the intersection result is determined, including: when the first positional relationship between the point of interest to be processed and the second interest surface is that the point of interest to be processed is inside the second interest surface, and the second positional relationship between the initial point of interest and the first interest surface is that the initial point of interest is inside the first interest surface, the intersection result indicates that the point of interest to be processed and the initial point of interest belong to the intersection region of the first interest surface and the second interest surface.

[0076] Specifically, since the interest point to be processed is within the first interest plane and the initial interest point is within the second interest plane, if the first positional relationship indicates that the interest point to be processed is within the second interest plane and the second positional relationship indicates that the initial interest point is within the first interest plane, then it is determined that the intersection result indicates that the interest point to be processed and the initial interest point belong to the intersection region of the first interest plane and the second interest plane.

[0077] It should be noted that if the interest planes of the candidate interest point and the interest point to be processed intersect and the candidate interest point and the interest point to be processed are within the intersection region, then it is highly likely that the interest point to be processed and the candidate interest point are duplicates.

[0078] Exemplarily, for the interest point to be processed POI-x and the initial interest point POI-y, POI-x belongs to the first interest plane AOI-x and the initial interest point POI-y belongs to the second interest plane AOI-y, as Figure 3 and Figure 4 shown, there is an intersection region between the first interest plane AOI-x and the second interest plane AOI-y, and both the interest point to be processed POI-x and the initial interest point POI-y are within the intersection region, indicating that the interest point to be processed POI-x and the initial interest point POI-y may be duplicates, and the initial interest point POI-y is used as the candidate interest point.

[0079] If the first positional relationship is that the interest point to be processed is within the second interest plane and the second positional relationship is that the initial interest point is not within the first interest plane, then it is determined that the intersection result indicates that there is an intersection region between the first interest plane and the second interest plane, but the interest point to be processed and the initial interest point are not both within the intersection region.

[0080] Or, if the first positional relationship is that the interest point to be processed is not within the second interest plane and the second positional relationship is that the initial interest point is within the first interest plane, then it is determined that the intersection result indicates that there is an intersection region between the first interest plane and the second interest plane, but the interest point to be processed and the initial interest point are not both within the intersection region.

[0081] Exemplarily, for the interest point to be processed POI-x and the initial interest point POI-y, POI-x belongs to the first interest plane AOI-x and the initial interest point POI-y belongs to the second interest plane AOI-y, as Figure 5 shown, there is no intersection region between the first interest plane AOI-x and the second interest plane AOI-y, as Figure 6 and Figure 7 shown, there is an intersection region between the first interest plane AOI-x and the second interest plane AOI-y, but the interest point to be processed POI-x and the initial interest point POI-y are not both within the intersection region, then it is less likely that the interest point to be processed POI-x and the initial interest point POI-y are duplicates, and the initial interest point is excluded.

[0082] In the above embodiments, according to the first positional relationship and the second positional relationship, the intersection result can be determined quickly and accurately, improving the efficiency of interest point duplicate checking.

[0083] Optionally, after obtaining the first interest plane where the interest point to be processed is located and the second interest plane where the initial interest point is located, it includes: when the first positional relationship between the interest point to be processed and the second interest plane indicates that the interest point to be processed is within the second interest plane, determining the intersection region between the first interest plane and the second interest plane, determining the third positional relationship between the initial interest point and the intersection region, and if the third positional relationship indicates that the initial interest point is within the intersection region, determining that the intersection result indicates that the interest point to be processed and the initial interest point belong to the intersection region of the first interest plane and the second interest plane.

[0084] Among them, determining the third positional relationship between the initial interest point and the intersection region can be: taking the initial interest point as the origin, determining a third ray in any direction, obtaining the third boundary line intersecting with the third ray among the boundary lines of the intersection region, and determining the third winding number according to the third boundary line and the first ray. When the third winding number belongs to the preset winding number interval, determining the third positional relationship as the initial interest point being within the intersection region.

[0085] Optionally, if the third winding number does not belong to the preset winding number interval, determining the third positional relationship as the initial interest point not being within the intersection region. Further, the intersection result indicates that there is an intersection region between the first interest plane and the second interest plane, the interest point to be processed belongs to the intersection region, and the initial interest point does not belong to the intersection region.

[0086] It should be noted that for the specific process of determining the third positional relationship between the initial interest point and the intersection region, reference can be made to the specific process of determining the first positional relationship between the interest point to be processed and the second interest plane in the above embodiments.

[0087] In the above embodiments, for the initially selected interest points, through the intersection result between the first interest plane where the interest point to be processed is located and the second interest plane where the initial interest point is located, further screening is performed among the initial interest points to obtain candidate interest points similar to the interest point to be processed, that is, candidate interest points that may be duplicates of the interest point to be processed. By reducing redundant data, the efficiency and accuracy of interest point duplicate checking are improved.

[0088] In some embodiments, taking the interest point to be processed and the candidate interest point as the centers respectively, intercepting the first interested image and the second interested image in the map, including: obtaining a first image containing the interest point to be processed and a second image containing the candidate interest point in the map; predicting the interested radius of the interest point to be processed and the first image through a radius prediction model to obtain a first interested radius; predicting the interested radius of the candidate interest point and the second image through the radius prediction model to obtain a second interested radius; taking the interest point to be processed as the center, intercepting the first interested image in the map according to the first interested radius; taking the candidate interest point as the center, intercepting the second interested image in the map according to the second interested radius.

[0089] Among them, the first image contains the interest point to be processed, and the second image contains the candidate interest point; in practical applications, taking the interest point to be processed as the center to determine a rectangular area with a default area to obtain the first image, and taking the candidate interest point as the center to determine a rectangular area with a default area to obtain the second image. The default area can be determined according to actual needs. For example, for all interest points in the map, the rectangular area corresponding to the default area can contain the relevant information of the interest point.

[0090] Specifically, the electronic device obtains a first image containing the interest point to be processed and a second image containing the candidate interest point in the map, inputs the first image and the interest point to be processed into the radius prediction model, analyzes the first image through the radius prediction model to determine the interested information around the interest point to be processed, determines the first interested radius, inputs the second image and the candidate interest point into the radius prediction model, analyzes the second image through the radius prediction model to determine the interested information around the candidate interest point, and determines the second interested radius.

[0091] As Figure 8 shown, the electronic device intercepts the first interested image ROI-x with a radius of the first interested radius centered on the interest point to be processed POI-x in the map, and intercepts the second interested image ROI-y with a radius of the second interested radius centered on the candidate interest point POI-y.

[0092] In the above embodiments, the first interested radius and the second interested radius are obtained by processing the first image and the second image through the radius prediction model, and the first interested image and the second interested image are intercepted in the map according to the first interested radius and the second interested radius respectively, so that the first interested image contains the interested information of the interest point to be processed, and the second interested image contains the interested information of the candidate interest point, thereby improving the quality of the first interested image and the second interested image, and also improving the accuracy of interest point duplicate judgment.

[0093] In some embodiments, the radius prediction model includes: a feature extraction unit and an interested radius determination unit; performing interested radius prediction on a point of interest to be processed and a first image through the radius prediction model to obtain a first interested radius, including: extracting features of the first image through the feature extraction unit to obtain first image features; performing interested radius prediction on the first image features and the point of interest to be processed through the interested radius determination unit to obtain the first interested radius;

[0094] Performing interested radius prediction on a candidate point of interest and a second image through the radius prediction model to obtain a second interested radius, including: extracting features of the second image through the feature extraction unit to obtain second image features; performing interested radius prediction on the second image features and the candidate point of interest through the interested radius determination unit to obtain the second interested radius.

[0095] Specifically, the electronic device inputs the first image into the feature extraction unit to obtain first image features, and the first image features may include the texture features and color distribution features of the first image. The first image features and the point of interest to be processed are input into the interested radius determination unit, so that the interested radius determination unit determines the interested features centered on the point of interest to be processed in the first image features, and then outputs the first interested radius.

[0096] The electronic device inputs the second image into the feature extraction unit to obtain second image features, and the second image features may include the texture image and color distribution features of the second image. The second image features and the candidate point of interest are input into the interested radius determination unit, so that the interested radius determination unit determines the interested features centered on the candidate point of interest in the second image features, and then outputs the second interested radius.

[0097] In some embodiments, the point of interest duplicate determination method further includes: obtaining a sample image including a first sample point of interest in a map; processing the sample image through a first initial model to obtain a predicted interested radius; adjusting the parameters of the first initial model according to the interested radius label corresponding to the first sample point of interest and the predicted interested radius until the first initial model converges to obtain a radius prediction model.

[0098] Wherein, the sample image includes the first sample point of interest; in practical applications, in the map, a rectangular area with a default area is determined centered on the first sample point of interest to obtain the sample image; the interested radius label is obtained by annotating the first sample point of interest in the sample image; the first initial model includes an initial feature extraction unit and an initial interested radius determination unit.

[0099] Specifically, the electronic device inputs the sample image into the initial feature extraction unit to obtain the feature image of the sample image, inputs the first sample interest point and the feature image of the sample image into the initial interested radius determination unit to obtain the predicted interested radius; substitutes the predicted interested radius and the interested radius label corresponding to the first sample interest point into the first preset loss function to obtain the first loss value, and adjusts the parameters of the first initial model (i.e., adjusts the parameters of the initial feature extraction unit and the initial interested radius determination unit) through the first loss value. Through the above process, the first initial model is iteratively trained until the first initial model converges to obtain the radius prediction model.

[0100] In the above embodiment, the first image and the second image are processed by the radius prediction model to obtain the first interested radius and the second interested radius, so that the first interested image contains the interested information of the interest point to be processed, and the second interested image contains the interested information of the candidate interest point, thereby improving the quality of the first interested image and the second interested image, and also improving the accuracy of interest point duplicate determination.

[0101] In some embodiments, the multi-modal network model performs duplicate determination processing on the first text information, the first interested image of the interest point to be processed, the second text information of the candidate interest point, and the second interested image to obtain a duplicate determination result, including: respectively extracting the features of the first text information and the second text information through the multi-modal network model to obtain the first text feature and the second text feature; respectively extracting the features of the first interested image and the second interested image to obtain the first image feature and the second image feature; performing feature fusion on the first text feature and the first image feature to obtain the first fusion feature; performing feature fusion on the second text feature and the second image feature to obtain the second fusion feature; and performing classification processing on the first fusion feature and the second fusion feature to obtain the duplicate determination result.

[0102] Among them, the multi-modal network model includes a text feature extraction unit, an image feature extraction unit, a fusion unit, and a classification unit; the text feature extraction unit is used to extract the features of the text information, the image feature extraction unit is used to extract the image features, the fusion unit is used to perform feature fusion, and the classification unit is used to determine the repetition probability and output the duplicate determination result.

[0103] Specifically, the electronic device inputs the first text information into the text feature extraction unit to obtain the first text feature, and inputs the second text information into the text feature extraction unit to obtain the second text feature; inputs the first image of interest into the image feature extraction unit to obtain the first image feature, and inputs the second image of interest into the image feature extraction unit to obtain the second image feature; inputs the first text feature and the first image feature into the fusion unit to obtain the first fusion feature, and inputs the second text feature and the second image feature into the fusion unit to obtain the second fusion feature; inputs the first fusion feature and the second fusion feature into the classification unit to obtain the repetition probability. If the repetition probability is greater than the probability threshold, it is determined that the duplicate detection result is a duplicate. If the repetition probability is not greater than the probability threshold, it is determined that the duplicate detection result is not a duplicate.

[0104] Among them, the probability threshold can be determined by searching the test set according to the low-dimensional unconstrained optimization algorithm (Powell algorithm).

[0105] In some embodiments, the method for duplicate detection of points of interest further includes: obtaining the first sample text information and the first sample image of interest of the second sample point of interest; obtaining the second sample text information and the second sample image of interest of the third sample point of interest; performing duplicate detection processing on the first sample text information, the first sample image of interest, the second sample text information, and the second sample image of interest through the second initial model to obtain a predicted duplicate detection result; adjusting the parameters of the second initial model according to the predicted duplicate detection result and the duplicate detection label between the second sample point of interest and the third sample point of interest until the second initial model converges to obtain a multi-modal network model.

[0106] Among them, the first sample text information is the text information of the second sample point of interest, and the first sample image of interest is an image intercepted from the map with the second sample point of interest as the center according to its radius of interest; the second sample text information is the text information of the third sample point of interest, and the second sample image of interest is an image intercepted from the map with the third sample point of interest as the center according to its radius of interest; the duplicate detection label represents the true value of whether there is a duplicate between the second sample point of interest and the third sample point of interest.

[0107] Specifically, the electronic device inputs the first sample text information, the first sample image of interest, the second sample text information, and the second sample image of interest into the second initial model, obtains the predicted duplicate detection result through the second initial model, substitutes the predicted duplicate detection result and the duplicate detection label into the second preset loss function to obtain the second loss value, adjusts the parameters of the second initial model through the second loss value, and iteratively trains the second initial model through the above process until the second initial model converges to obtain a multi-modal network model.

[0108] In the above embodiments, the first text information, the first interested image of the interest point to be processed, the second text information of the candidate interest point, and the second interested image are subjected to duplicate checking processing through a multimodal network model to obtain a duplicate checking result. By fusing the text and image data of the interest point to be processed and the candidate interest point, the richness of the data used for interest point duplicate checking is improved, and the accuracy of interest point duplicate checking is improved.

[0109] Figure 9 Flow schematic of the interest point duplicate checking method provided by this application Figure 2 , as Figure 9 shown, on the basis of the Figure 2 embodiment, the interest point duplicate checking method is described in detail. The method includes:

[0110] S901. In response to a duplicate checking instruction for the interest point to be processed, recall the map interest points based on the interest point to be processed to screen the initial interest points in the map interest points;

[0111] S902. Obtain the first interest surface where the interest point to be processed is located and the second interest surface where the initial interest point is located;

[0112] S903. Taking the interest point to be processed as the origin, determine a first ray in any direction; obtain a first boundary line intersecting the first ray among the boundary lines of the second interest surface; determine a first winding number based on the first ray and the first boundary line; when the first winding number belongs to a preset winding number interval, determine that the first positional relationship between the interest point to be processed and the second interest surface is that the interest point to be processed is inside the second interest surface;

[0113] S904. Taking the initial interest point as the origin, determine a second ray in any direction; obtain second boundary lines intersecting the second ray among the boundary lines of the first interest surface; determine a second winding number based on the second ray and the second boundary lines; when the second winding number belongs to a preset winding number interval, determine that the second positional relationship between the initial interest point and the first interest surface is that the initial interest point is inside the first interest surface;

[0114] S905. When the first positional relationship is that the interest point to be processed is inside the second interest surface and the second positional relationship is that the initial interest point is inside the first interest surface, the intersection result indicates that the interest point to be processed and the initial interest point belong to the intersection area of the first interest surface and the second interest surface, and the initial interest point is used as a candidate interest point similar to the interest point to be processed;

[0115] S906. Obtain a first image containing the interest point to be processed and a second image containing candidate interest points in the map; perform an interested radius prediction on the interest point to be processed and the first image through a radius prediction model to obtain a first interested radius; perform an interested radius prediction on the candidate interest point and the second image through the radius prediction model to obtain a second interested radius; take the interest point to be processed as the center and intercept a first interested image in the map according to the first interested radius; take the candidate interest point as the center and intercept a second interested image in the map according to the second interested radius.

[0116] S907. Through a multimodal network model, perform feature extraction on the first text information and the second text information respectively to obtain a first text feature and a second text feature; perform feature extraction on the first interested image and the second interested image respectively to obtain a first image feature and a second image feature; perform feature fusion on the first text feature and the first image feature to obtain a first fusion feature; through the fusion unit of the multimodal network model, perform feature fusion on the second text feature and the second image feature to obtain a second fusion feature; perform classification processing on the first fusion feature and the second fusion feature to obtain a duplicate determination result.

[0117] The interest point duplicate determination method provided by the embodiments of this application screens out candidate interest points similar to the interest point to be processed among the map interest points, intercepts a first interested image centered on the interest point to be processed and a second interested image centered on the candidate interest point in the map, introduces high-order visual features around the interest point to be processed and the candidate interest point, and performs duplicate determination processing on the first text information, the first interested image of the interest point to be processed, the second text information, and the second interested image of the candidate interest point through a multimodal network model to obtain a duplicate determination result. By fusing the text and image data of the interest point to be processed and the candidate interest point, the richness of the data used for interest point duplicate determination is improved, and the accuracy of interest point duplicate determination is improved.

[0118] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the arrows, these steps do not necessarily need to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily need to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily need to be sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0119] Figure 10The following is a schematic structural diagram of the point of interest duplication determination device provided by this application. As Figure 10 shown, the point of interest duplication determination device 100 provided in this embodiment includes:

[0120] A selection module 1001, configured to, in response to a duplication determination instruction for a to-be-processed point of interest, select candidate points of interest similar to the to-be-processed point of interest from the map points of interest;

[0121] An image cropping module 1002, configured to respectively center on the to-be-processed point of interest and the candidate point of interest, and crop a first region of interest image and a second region of interest image in the map;

[0122] A multi-modal processing module 1003, configured to perform duplication determination processing on the first text information, the first region of interest image of the to-be-processed point of interest, the second text information, and the second region of interest image of the candidate point of interest through a multi-modal network model, and obtain a duplication determination result, where the duplication determination result is used to indicate whether the to-be-processed point of interest and the candidate point of interest are duplicates.

[0123] In a possible implementation manner, the selection module 1001 is further configured to perform a recall process on the map points of interest according to the to-be-processed point of interest, so as to screen initial points of interest in the map points of interest; obtain a first region of interest where the to-be-processed point of interest is located and a second region of interest where the initial point of interest is located; determine an intersection result according to a first positional relationship between the to-be-processed point of interest and the second region of interest, and a second positional relationship between the initial point of interest and the first region of interest; when the intersection result indicates that the to-be-processed point of interest and the initial point of interest belong to the intersection region of the first region of interest and the second region of interest, use the initial point of interest as a candidate point of interest similar to the to-be-processed point of interest.

[0124] In a possible implementation manner, the selection module 1001 is further configured to use the to-be-processed point of interest as the origin, and determine a first ray in any direction; obtain a first boundary line intersecting with the first ray in the boundary line of the second region of interest; determine a first winding number according to the first ray and the first boundary line; when the first winding number belongs to a preset winding number interval, determine that the first positional relationship between the to-be-processed point of interest and the second region of interest is that the to-be-processed point of interest is inside the second region of interest.

[0125] In a possible implementation manner, the selection module 1001 is further configured to use the initial point of interest as the origin, and determine a second ray in any direction; obtain a second boundary line intersecting with the second ray in each boundary line of the first region of interest; determine a second winding number according to the second ray and the second boundary line; when the second winding number belongs to a preset winding number interval, determine that the second positional relationship between the initial point of interest and the first region of interest is that the initial point of interest is inside the first region of interest.

[0126] In a possible implementation, the selection module 1001 is further configured to, when the first positional relationship is that the interest point to be processed is inside the second interest surface and the second positional relationship is that the initial interest point is inside the first interest surface, the intersection result indicates that the interest point to be processed and the initial interest point belong to the intersection area of the first interest surface and the second interest surface.

[0127] In a possible implementation, the image capture module 1002 is further configured to obtain a first image including the interest point to be processed and a second image including the candidate interest point in the map; perform an interested radius prediction on the interest point to be processed and the first image through a radius prediction model to obtain a first interested radius; perform an interested radius prediction on the candidate interest point and the second image through the radius prediction model to obtain a second interested radius; take the interest point to be processed as the center and intercept a first interested image in the map according to the first interested radius; take the candidate interest point as the center and intercept a second interested image in the map according to the second interested radius.

[0128] In a possible implementation, the radius prediction model includes: a feature extraction unit and an interested radius determination unit; the image capture module 1002 is further configured to extract features from the first image through the feature extraction unit to obtain first image features; perform an interested radius prediction on the first image features and the interest point to be processed through the interested radius determination unit to obtain a first interested radius;

[0129] Extract features from the second image through the feature extraction unit to obtain second image features; perform an interested radius prediction on the second image features and the candidate interest point through the interested radius determination unit to obtain a second interested radius.

[0130] In a possible implementation, the multi-modal processing module 1003 is further configured to extract features from the first text information and the second text information respectively through a multi-modal network model to obtain first text features and second text features; extract features from the first interested image and the second interested image respectively to obtain first image features and second image features; perform feature fusion on the first text features and the first image features to obtain first fusion features; perform feature fusion on the second text features and the second image features to obtain second fusion features; perform classification processing on the first fusion features and the second fusion features to obtain a duplicate determination result.

[0131] The interest point duplicate determination device provided in this embodiment can execute the interest point duplicate determination method provided in the above method embodiment, and its implementation principle and technical effects are similar, which will not be elaborated here in this embodiment.

[0132] Figure 11 It is a schematic structural diagram of an electronic device provided in this application. As Figure 11As shown in the figure, the electronic device 110 provided in this embodiment includes: at least one processor 1101 and a memory 1102. Optionally, the device 110 further includes a communication component 1103. Among them, the processor 1101, the memory 1102, and the communication component 1103 are connected through a bus.

[0133] In the specific implementation process, at least one processor 1101 executes the computer-executable instructions stored in the memory 1102, so that at least one processor 1101 executes the above-mentioned method.

[0134] For the specific implementation process of the processor 1101, reference can be made to the above method embodiment. The implementation principle and technical effect are similar, and will not be elaborated here in this embodiment.

[0135] In the above embodiment, it should be understood that the processor may be a central processing unit (English: Central Processing Unit, abbreviated: CPU), or other general-purpose processors, digital signal processors (English: Digital Signal Processor, abbreviated: DSP), application specific integrated circuits (English: Application Specific Integrated Circuit, abbreviated: ASIC), etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the invention can be directly embodied as being executed by a hardware processor, or executed by a combination of hardware and software modules in the processor.

[0136] The memory may include a high-speed memory (Random Access Memory, RAM), and may also include a non-volatile memory (Non-volatile Memory, NVM), such as at least one disk memory.

[0137] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, the buses in the drawings of this application are not limited to only one bus or one type of bus.

[0138] This application also provides a computer program product, including a computer program, which implements the above-mentioned method when executed by a processor.

[0139] The present application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0140] The above-readable storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disk. The readable storage medium may be any available medium accessible by a general-purpose or special-purpose computer.

[0141] An exemplary readable storage medium is coupled to the processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium may also be an integral part of the processor. The processor and the readable storage medium may be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium may also exist as discrete components in a device.

[0142] The division of units is merely a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the couplings or direct couplings or communication connections shown or discussed among each other may be indirect couplings or communication connections through some interfaces, devices, or units, and may be in electrical, mechanical, or other forms.

[0143] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, i.e., they may be located in one place or distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0144] Furthermore, in each embodiment of the present invention, the functional units may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit.

[0145] If a function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs, etc., all kinds of media that can store program codes.

[0146] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When this program is executed, it executes the steps including the above method embodiments; and the aforementioned storage medium includes: ROMs, RAMs, magnetic disks, or optical discs, etc., all kinds of media that can store program codes.

[0147] Finally, it should be noted that: After considering the specification and practicing the invention disclosed herein, those skilled in the art will easily think of other implementation manners of the present invention. The present invention is intended to cover any variations, uses, or adaptive changes of the present invention. These variations, uses, or adaptive changes follow the general principles of the present invention and include the common general knowledge or conventional technical means in the technical field not disclosed in the present invention. It is not limited to the exact structures already described and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present invention is only limited by the appended claims.

Claims

1. A method for determining duplicate points of interest, characterized in that: include: In response to a duplicate determination instruction for a pending interest point, selecting a candidate interest point similar to the pending interest point from among the map interest points; Taking the to-be-processed interest point and the candidate interest point as the center respectively, intercepting a first image of interest and a second image of interest in the map; The first text information of the interest point to be processed, the first image of interest, the second text information of the candidate interest point and the second image of interest are subjected to duplicate detection processing through a multimodal network model to obtain a duplicate detection result, wherein the duplicate detection result is used to indicate whether the interest point to be processed and the candidate interest point are repeated.

2. The method according to claim 1, characterized in that The selecting of candidate interest points similar to the interest point to be processed from the interest points on the map includes: Recalling the map interest points according to the pending interest points to select initial interest points from the map interest points; Acquire a first interest surface where the to-be-processed interest point is located and a second interest surface where the initial interest point is located; Determining an intersection result according to a first position relationship between the interest point to be processed and the second interest surface, and a second position relationship between the initial interest point and the first interest surface; When the intersection result indicates that the interest point to be processed and the initial interest point belong to the intersection area of ​​the first interest plane and the second interest plane, the initial interest point is used as a candidate interest point similar to the interest point to be processed.

3. The method according to claim 2, characterized in that After obtaining the first interest plane where the to-be-processed interest point is located and the second interest plane where the initial interest point is located, the method further includes: Taking the interest point to be processed as the origin, determining a first ray in any direction; Acquire a first boundary line intersecting with the first ray from the boundary line of the second interest plane; determining a first winding number according to the first ray and the first boundary line; When the first surround number belongs to a preset surround number interval, it is determined that the first position relationship between the interest point to be processed and the second interest surface is that the interest point to be processed is inside the second interest surface.

4. The method according to claim 2, characterized in that: After obtaining the first interest surface where the to-be-processed interest point is located and the second interest surface where the initial interest point is located, the method further includes: Taking the initial point of interest as the origin, determining a second ray in any direction; Acquire a second boundary line intersecting with the second ray from each boundary line of the first interest surface; determining a second winding number according to the second ray and the second boundary line; When the second winding number belongs to a preset winding number interval, it is determined that the second positional relationship between the initial interest point and the first interest surface is that the initial interest point is inside the first interest surface.

5. The method according to claim 2, characterized in that: The determining the intersection result according to the first position relationship between the to-be-processed interest point and the second interest surface, and the second position relationship between the initial interest point and the first interest surface, comprises: When the first position relationship between the to-be-processed interest point and the second interest surface is that the to-be-processed interest point is inside the second interest surface, and the second position relationship between the initial interest point and the first interest surface is that the initial interest point is inside the first interest surface, the intersection result indicates that the to-be-processed interest point and the initial interest point belong to the intersection area of ​​the first interest surface and the second interest surface.

6. The method according to any one of claims 1 to 5, characterized in that The method of capturing a first image of interest and a second image of interest in a map with the to-be-processed point of interest and the candidate point of interest as the center respectively includes: Acquire a first image containing the interest point to be processed and a second image containing the candidate interest point in a map; Predicting the radius of interest of the to-be-processed interest point and the first image by using a radius prediction model to obtain a first radius of interest; Predicting the radius of interest for the candidate interest point and the second image using the radius prediction model to obtain a second radius of interest; Taking the to-be-processed point of interest as the center, intercepting a first image of interest in the map according to the first radius of interest; Taking the candidate point of interest as the center, a second image of interest is captured in the map according to the second radius of interest.

7. The method according to any one of claims 1 to 5, characterized in that The method of performing duplicate detection processing on the first text information of the interest point to be processed, the first image of interest, the second text information of the candidate interest point, and the second image of interest by using the multimodal network model to obtain a duplicate detection result includes: Extract features from the first text information and the second text information respectively through a multimodal network model to obtain first text features and second text features; Extracting features from the first image of interest and the second image of interest respectively to obtain first image features and second image features; Performing feature fusion on the first text feature and the first image feature to obtain a first fused feature; Performing feature fusion on the second text feature and the second image feature to obtain a second fused feature; The first fusion feature and the second fusion feature are classified and processed to obtain a duplicate detection result.

8. A device for determining duplicate points of interest, characterized in that: The device comprises: A selection module, configured to select, in response to a duplicate determination instruction for a pending interest point, a candidate interest point similar to the pending interest point from among the map interest points; An image capture module, used to capture a first image of interest and a second image of interest in a map with the to-be-processed point of interest and the candidate point of interest as the center respectively; The multimodal processing module is used to perform duplicate detection processing on the first text information of the interest point to be processed, the first image of interest, the second text information of the candidate interest point and the second image of interest through a multimodal network model to obtain a duplicate detection result, wherein the duplicate detection result is used to indicate whether the interest point to be processed and the candidate interest point are repeated.

9. An electronic device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium / computer program product, characterized in that: The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor; and / or, The computer program product comprises computer-executable instructions, and when the computer-executable instructions are executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images