POI (Point of Interest) data processing method and device, server, medium and program product

Abstract

The invention provides an interest point data processing method and device, a server, a medium and a program product, and relates to the technical field of electronic maps, the method comprises the following steps: obtaining first error POI information of a position POI according to a position coordinate of a first geographic element carried in a name of the position POI and first difference information between the position coordinate of the position POI; wherein the first geographic element is an element obtained by performing name segmentation on the position POI; acquiring second error POI information of the position POI according to second difference information between position coordinates of a second geographic element carried in the address information of the position POI and the position coordinates of the position POI; wherein the second geographic element is an element obtained by performing address segmentation on the position POI; and determining the position POI as an error POI according to at least one of the first error POI information and the second error POI information. By means of the method, the recognition efficiency and precision of the wrong POI can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of electronic maps, and in particular, to a method, device, server, medium, and program product for processing point of interest data. Background Art

[0002] In an electronic map, in order to facilitate a user to quickly locate a target position, the target position is usually marked in the form of a point of interest (POI).

[0003] However, with the continuous update and expansion of electronic map data, the accuracy and reliability problems of location POI data gradually emerge. For example, due to various problems such as data input errors, untimely information updates, and unreflected changes in merchant locations, there will be a large number of incorrect location POIs in the system, resulting in the user's difficulty in quickly locating the corresponding location points and affecting the user's map usage experience.

[0004] Therefore, there is an urgent need to propose a technical solution that can effectively screen incorrect location POIs to solve the above problems. Summary of the Invention

[0005] The present application provides a method, device, server, medium, and program product for processing point of interest data to at least solve one of the above technical problems.

[0006] According to one aspect of the present application, there is provided a method for processing point of interest data, including: obtaining first incorrect POI information of the location POI according to first difference information between position coordinates of a first geographical element carried in a name of the location point of interest (POI) and the position coordinates of the location POI; wherein, the first geographical element is an element obtained by splitting the name of the location POI; obtaining second incorrect POI information of the location POI according to second difference information between position coordinates of a second geographical element carried in address information of the location POI and the position coordinates of the location POI; wherein, the second geographical element is an element obtained by splitting the address of the location POI; determining the location POI as an incorrect POI according to at least one of the first incorrect POI information and the second incorrect POI information.

[0007] In one embodiment, the name splitting of the location POI includes: splitting the name of the location POI according to a preset name splitting model to obtain a name splitting result, where the name splitting result includes at least one of the first geographical elements; wherein, the name splitting model is trained based on historical location POI data and is used to split and identify the first geographical elements carried in the name of the location POI; and / or, the address splitting of the location POI includes: splitting and identifying the address of the location POI according to a preset address splitting model to obtain an address splitting result, where the address splitting result includes at least one of the second geographical elements; wherein, the address splitting model is trained based on historical location POI data and is used to split and identify the second geographical elements carried in the address of the location POI.

[0008] In one embodiment, the method further includes: obtaining a first element type of the first geographical element according to a preset name splitting model, where the first element type includes at least one of the following: road element, landmark element, and geographical interest area element; the obtaining of the first incorrect POI information of the location POI according to the first difference information between the position coordinates of the first geographical element carried in the name of the location of interest point POI and the position coordinates of the location POI includes: obtaining the first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographical element and the position coordinates of the location POI.

[0009] In one embodiment, the first difference information includes a distance difference and / or cross-road information; the obtaining of the first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographical element and the position coordinates of the location POI includes: if the first geographical element is a road element, obtaining the distance difference between the position coordinates of the road element and the position coordinates of the location POI, and when the distance difference reaches a first preset distance threshold, obtaining the first incorrect POI information of the location POI regarding the road distance; and / or, obtaining the cross-road information between the position coordinates of the road element and the position coordinates of the location POI, where the cross-road information is used to indicate whether there are other roads other than the road element in the straight-line area connecting the position coordinates of the road element and the position coordinates of the location POI, and if so, obtaining the first incorrect POI information of the location POI regarding the cross-road.

[0010] In one embodiment, the first difference information includes a distance difference. The method for obtaining the first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographic element and the position coordinates of the location POI includes: if the first geographic element is a landmark element, obtaining the distance difference between the position coordinates of the landmark element and the position coordinates of the location POI, and when the distance difference reaches a second preset distance threshold, obtaining the first incorrect POI information of the location POI regarding the landmark distance; wherein, the second preset distance threshold is less than or equal to the first preset distance threshold.

[0011] In one embodiment, the first difference information includes cross-interest area information; the method for obtaining the first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographic element and the position coordinates of the location POI includes: if the first geographic element is a geographic interest area element, obtaining the cross-interest area information between the position coordinates of the geographic interest area element and the position coordinates of the location POI, where the cross-interest area information is used to indicate whether the position coordinates of the location POI are within the position coordinates of the geographic interest area element, and if not, obtaining the first incorrect POI information of the location POI regarding the cross-interest area.

[0012] In one embodiment, the training method of the name segmentation model includes: collecting historical POI data, and performing annotation processing on the first geographic elements carried in the names in the historical POI data to obtain a first training sample set; and based on the first training sample set, training a deep learning model for sequence labeling tasks to obtain the name segmentation model; and / or, the training method of the address segmentation model includes: collecting historical POI data, and performing annotation processing on the second geographic elements carried in the addresses in the historical POI data to obtain a second training sample set, and training a deep learning model for sequence labeling tasks to obtain the address segmentation model.

[0013] In one embodiment, the method further includes: obtaining a second element type of the second geographical element according to a preset address segmentation model, where the second element type includes at least one of administrative division elements and first element type elements; obtaining second incorrect POI information of the location POI according to second difference information between the position coordinates of the second geographical element carried in the address information of the location POI and the position coordinates of the location POI, including: if the second geographical element is an administrative division element, sequentially determining whether the position coordinates of the location POI are within the position coordinates of the administrative division element at each level of administrative division range, and if not, obtaining second incorrect POI information of the location POI regarding the administrative division; and / or, if the second geographical element is a first element type element, obtaining second incorrect POI information of the location POI regarding at least one of roads, landmarks, and cross-interest regions according to the corresponding difference information between the second geographical element and the location POI.

[0014] In one embodiment, the location POI is any POI in the POI database; the method further includes: screening out all location POIs determined to be incorrect POIs from the POI database.

[0015] In one embodiment, the method further includes: obtaining address segmentation results of all location POIs in the POI database according to the address segmentation model to obtain respective second geographical elements of each location POI; clustering the location POIs with the same second geographical element to obtain a clustering result; calculating position distances between respective clustered location POIs in the clustering result, and screening out incorrect POIs with position distances greater than a third preset threshold from the clustered location POIs according to the position distance results.

[0016] According to a second aspect of the present application, there is provided a point of interest data processing apparatus, including: a first recognition module configured to obtain first incorrect POI information of the location POI according to first difference information between the position coordinates of a first geographical element carried in the name of the location point of interest POI and the position coordinates of the location POI; where the first geographical element is an element obtained by segmenting the name of the location POI; a second recognition module configured to obtain second incorrect POI information of the location POI according to second difference information between the position coordinates of a second geographical element carried in the address information of the location POI and the position coordinates of the location POI; where the second geographical element is an element obtained by segmenting the address of the location POI; an incorrect POI determination module configured to determine the location POI as an incorrect POI according to at least one of the first incorrect POI information and the second incorrect POI information.

[0017] In one embodiment, the device further includes a name segmentation module for segmenting the name of the location POI. The name segmentation module is configured to segment the name of the location POI according to a preset name segmentation model, and obtain a name segmentation result, where the name segmentation result includes at least one of the first geographical elements; wherein, the name segmentation model is trained based on historical location POI data and is used to segment and identify the first geographical elements carried by the name in the location POI.

[0018] In one embodiment, the device further includes an address segmentation module for segmenting the address of the location POI. The address segmentation module is configured to segment and identify the address of the location POI according to a preset address segmentation model, and obtain an address segmentation result, where the address segmentation result includes at least one of the second geographical elements; wherein, the address segmentation model is trained based on historical location POI data and is used to segment and identify the second geographical elements carried by the address in the location POI.

[0019] In one embodiment, the device further includes: a first element type acquisition module configured to obtain a first element type of the first geographical element according to a preset name segmentation model, where the first element type includes at least one of the following: road element, landmark element, and geographical interest area element; the first identification module is specifically configured to obtain first wrong POI information of the location POI according to the first element type and first difference information between the position coordinates of the first geographical element and the position coordinates of the location POI.

[0020] In one embodiment, the first difference information includes a distance difference and / or cross-road information; the first identification module is specifically configured to: if the first geographical element is a road element, obtain a distance difference between the position coordinates of the road element and the position coordinates of the location POI, and when the distance difference reaches a first preset distance threshold, obtain first wrong POI information of the location POI regarding the road distance; and / or, obtain cross-road information between the position coordinates of the road element and the position coordinates of the location POI, where the cross-road information is used to indicate whether there are other roads except the road element in the straight-line area connecting the position coordinates of the road element and the position coordinates of the location POI. If so, obtain first wrong POI information of the location POI regarding cross-road.

[0021] In one embodiment, the first difference information includes a distance difference; the first recognition module is further configured to: if the first geographical element is a landmark element, obtain the distance difference between the position coordinates of the landmark element and the position coordinates of the position POI, and when the distance difference reaches a second preset distance threshold, obtain the first incorrect POI information of the position POI regarding the landmark distance; wherein, the second preset distance threshold is less than or equal to the first preset distance threshold.

[0022] In one embodiment, the first difference information includes cross-interest region information; the first recognition module is specifically configured to: if the first geographical element is a geographical interest region element, obtain the cross-interest region information between the position coordinates of the geographical interest region element and the position coordinates of the position POI, where the cross-interest region information is used to indicate whether the position coordinates of the position POI are within the position coordinates of the geographical interest region element, and if not, obtain the first incorrect POI information of the position POI regarding the cross-interest region.

[0023] In one embodiment, it further includes a first training module for training the name segmentation model, which is configured to collect historical POI data, perform annotation processing on the first geographical features carried by the names in the historical POI data to obtain a first training sample set; and based on the first training sample set, train a deep learning model for sequence labeling tasks to obtain the name segmentation model;

[0024] In one embodiment, it further includes a second training module for training the address segmentation model, which is configured to collect historical POI data, perform annotation processing on the second geographical features carried by the addresses in the historical POI data to obtain a second training sample set, and train a deep learning model for sequence labeling tasks to obtain the address segmentation model.

[0025] In one embodiment, the device further includes: a second element type acquisition module, which is configured to obtain the second element type of the second geographical element according to a preset address segmentation model, where the second element type includes at least one of administrative division elements and first element type elements; the second recognition module is specifically configured to: if the second geographical element is an administrative division element, sequentially determine whether the position coordinates of the position POI are within the administrative division ranges at all levels corresponding to the position coordinates of the administrative division element, and if not, obtain the second incorrect POI information of the position POI regarding the administrative division; and / or, if the second geographical element is a first element type element, obtain the second incorrect POI information of the position POI regarding at least one of roads, landmarks, and cross-interest regions according to the corresponding difference information between the second geographical element and the position POI.

[0026] In one embodiment, the location POI is any POI in the POI database; the apparatus further includes: an incorrect POI screening module configured to screen out all location POIs determined to be incorrect POIs from the POI database.

[0027] In one embodiment, the apparatus further includes: a segmentation acquisition module configured to obtain the address segmentation results of all location POIs in the POI database according to an address segmentation model to obtain respective second geographical elements of each location POI; a clustering module configured to cluster the location POIs having the same second geographical element to obtain a clustering result; the incorrect POI screening module is further configured to calculate the location distances between the clustered location POIs in the clustering result, and screen out incorrect POIs with a location distance greater than a third preset threshold from the clustered location POIs according to the location distance result.

[0028] According to a third aspect of the present application, there is provided a server, the server including: a memory and a processor; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory, so that the server executes the point of interest data processing method provided in any one of the above first aspects.

[0029] According to a fourth aspect of the present application, there is provided a computer-readable storage medium storing computer execution instructions, and the computer execution instructions are used to implement the point of interest data processing method provided in any one of the above first aspects when executed by a processor.

[0030] According to a fifth aspect of the present application, there is provided a computer program product including a computer program, and the computer program implements the point of interest data processing method provided in any one of the above first aspects when executed by a processor.

[0031] The method, apparatus, server, medium and program product for processing point of interest (POI) data provided by this application obtain the first incorrect POI information of a location POI based on the first difference information between the position coordinates of the first geographical element carried in the name of the location POI and the position coordinates of the location POI. The first geographical element is an element obtained by splitting the name of the location POI. The second incorrect POI information of the location POI is obtained based on the second difference information between the position coordinates of the second geographical element carried in the address information of the location POI and the position coordinates of the location POI. The second geographical element is an element obtained by splitting the address of the location POI. Based on at least one of the first incorrect POI information and the second incorrect POI information, it is determined that the location POI is an incorrect POI. In this process, by splitting the name and address of the location POI, the geographical (address) elements therein are identified, and combined with the difference information between the position coordinates of the location POI, so as to quickly identify the incorrect data in the mismatch situation existing in the POI address - POI name - POI coordinate triple. This process can automatically detect incorrect POIs in a large number of POI data without manual participation, reducing the labor cost, with higher identification efficiency and no missed identification, facilitating the quick screening of incorrect POIs in the POI library, thus solving the problem that users are difficult to quickly locate the corresponding position points, affecting the user experience of using the map. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings herein are incorporated into and constitute a part of this specification, showing embodiments consistent with this application, and are used together with the specification to explain the principles of this application.

[0033] Figure 1 It is a schematic diagram of a possible scenario provided by an embodiment of this application;

[0034] Figure 2 It is a schematic flowchart of a method for processing point of interest (POI) data provided by an embodiment of this application;

[0035] Figure 3 It is a schematic flowchart of another method for processing point of interest (POI) data provided by an embodiment of this application;

[0036] Figure 4 It is a schematic flowchart of yet another method for processing point of interest (POI) data provided by an embodiment of this application;

[0037] Figure 5 It is an example diagram of a location POI crossing a road in an embodiment of this application;

[0038] Figure 6 It is an example diagram of the AOI range in an embodiment of this application;

[0039] Figure 7A schematic flowchart of another method for processing point-of-interest data provided by an exemplary embodiment of the present application;

[0040] Figure 8 A schematic flowchart of yet another method for processing point-of-interest data provided by an embodiment of the present application;

[0041] Figure 9a A schematic flowchart of a process for identifying incorrect POIs through clustering in an embodiment of the present application;

[0042] Figure 9b An example diagram of an outlier POI in an embodiment of the present application;

[0043] Figure 10 A schematic structural diagram of a point-of-interest data processing device provided by an embodiment of the present application;

[0044] Figure 11 A schematic structural diagram of a server provided by an embodiment of the present application.

[0045] Through the above-mentioned drawings, specific embodiments of the present 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 the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. Detailed implementation manners

[0046] 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 implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0047] The embodiments of the present application will be explained below in combination with an application scenario. The method for processing point-of-interest data provided by the embodiments of the present application can be applied to the application scenario of map annotation. Exemplarily, the execution subject of the method provided by the embodiments of the present application can be a server. More specifically, for example, it can be the server of a location POI data mining manufacturer. The following will introduce the method provided by the embodiments of the present application with the server as the execution subject.

[0048] Figure 1 A schematic diagram of the scenario of a method for processing point-of-interest data provided by an embodiment of the present application, as Figure 1As shown in the figure, the server 110 and the POI database 120 are electrically connected. The POI database is used to store location POI data, which can be a relational database or a non-relational database. This embodiment does not make any special limitation on this. The server 110 can extract target location POI data from the POI database through SQL (Structured Query Language) queries or API (Application Programming Interface) requests. For example, data can be extracted according to the name, location, type, update time, etc. of the POI. Alternatively, as needed, the server 110 can extract a large amount of data at one time for batch processing, or extract subset data according to specific conditions for location POI data processing, and screen out the data determined to be incorrect POI. Optionally, the POI database 120 is also electrically connected to the intelligent vehicle 130. When the intelligent vehicle needs to use an electronic map (such as a high-precision map) for assisted driving, it uses the API of the map server 140 to load the basic electronic map, and adds annotations (markers) on the map according to the POI data in the POI database 120, combined with the POI data name or address description. For example, different types of location POIs can be represented by icons or symbols, so as to facilitate users to quickly locate according to the location POI. In some application scenarios, the above method for screening incorrect POI can also be integrated into the POI database 120. Alternatively, the POI database 120 can be connected to the map server 140, and the map server performs POI annotation on the map and then publishes it to the intelligent vehicle for use. This embodiment does not make any special limitation on the specific application method of the POI data. The server 110 can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, and cloud computing.

[0049] In the above process, for the discovery of incorrect POI data in the POI database, currently it is mainly achieved through user feedback and manual verification. For a large amount of POI data, using the method of manual verification, the labor cost is relatively high, it is difficult to achieve large-scale manual verification, and since most of the POI data is correct, manual verification is mostly for correct data in most cases, wasting a large amount of human resources. Moreover, during the manual verification process, it is often difficult to verify due to the lack of corresponding verification materials (such as POI address pictures, etc.), and the verification accuracy is relatively low. Relying on user feedback, it cannot be guaranteed that incorrect POI will be reported by users, and a large amount of low-popularity incorrect POI data remains in the POI database and cannot be cleaned.

[0050] In view of this, an embodiment of the present application provides a method, apparatus, server, medium and program product for processing point of interest (POI) data. By obtaining the first incorrect POI information of a location POI based on the first difference information between the position coordinates of a first geographical element carried in the name of the location POI, where the first geographical element is an element obtained by splitting the name of the location POI, and the position coordinates of the location POI, and obtaining the second incorrect POI information of the location POI based on the second difference information between the position coordinates of a second geographical element carried in the address information of the location POI, where the second geographical element is an element obtained by splitting the address of the location POI, and determining that the location POI is an incorrect POI based on at least one of the first incorrect POI information and the second incorrect POI information. In this process, by splitting the name and address of the location POI, the geographical (address) elements therein are identified, and combined with the difference information from the position coordinates of the location POI, to quickly determine the incorrect data of the mismatch situation existing in the POI address - POI name - POI coordinate triple. This process can automatically detect incorrect POIs in a large number of POI data without manual participation, reducing labor costs, with higher identification efficiency and no missed identification, facilitating the rapid screening of incorrect POIs in the POI library, thus solving the problem that users are difficult to quickly locate the corresponding location points, which affects the user's map usage experience.

[0051] The following will specifically describe the technical solutions of the present application and how the technical solutions of the present application solve the above technical problems in conjunction with the accompanying drawings and specific embodiments. It should be noted that these specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0052] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards, and corresponding operation entrances are provided for users to choose to authorize or refuse.

[0053] Figure 2 is a flowchart of a method for processing point of interest data provided by an embodiment of the present application. The execution entity can be Figure 1 a server in the application scenario, or it can also be a POI database. This embodiment does not make a special limitation on the specific execution entity of this method. As shown Figure 2 below, this method includes steps S201 - S203:

[0054] Step S201: Obtain the first incorrect POI information of the location POI according to the first difference information between the position coordinates of the first geographical element carried in the name of the location point of interest (POI) and the position coordinates of the location POI; wherein, the first geographical element is an element obtained by splitting the name of the location POI.

[0055] Location POIs can include commercial facilities such as restaurants, hotels, shopping malls, etc., or public service facilities such as hospitals, schools, subway stations, etc. Marking location POIs on the map enables users to quickly find the required location information through simple searches or map browsing.

[0056] When a location POI is published or uploaded, it carries the name and address of the location POI. The name of the location POI is a text label used to identify the location POI, usually the official name or common name of the location. In addition, considering the coincidence of the same store names, the name usually also carries branch or main store information to facilitate user quick positioning. Examples are as follows: Store A (XXX Building Store), Store B (XXX Road), where XXX Building and XXX Road are the first geographical elements carried in the location POI. Considering that the name of the location POI may carry multiple geographical elements and other non-geographical element words, to facilitate the identification of geographical elements in the name, the name of the location POI is split to obtain the first geographical element. The name splitting method can use a machine learning model such as a natural language model for splitting and identification, or it can also adopt the method of rule matching (i.e., predefined rules and patterns (such as regular expressions) to identify and extract geographical elements, which can be determined by the user according to prior data) for splitting and identification. This embodiment does not make special limitations on this.

[0057] The position coordinates of the first geographical element, such as the geographical location coordinates of XXX Building, can be obtained through real-time data acquisition methods. Exemplarily, the acquisition method of the position coordinates of the first geographical element can be obtained from a geocoding system, which is a process and technology that converts an address or location description into geographical coordinates (latitude and longitude), and can be constructed using a large number of geographical elements and their corresponding geographical coordinates. The geocoding system can quickly query the position coordinates corresponding to the geographical elements. In some examples, the position coordinates of geographical elements can also be obtained through other methods, such as the Global Positioning System (GPS), Geographic Information System (GIS), etc.

[0058] The location coordinates of the location POI, that is, the actual location coordinates of the location POI. For example, the location coordinates of store A (store in XXX Building) can also be obtained through the above methods, such as using the geocoding system or the GPS system, etc.

[0059] The first difference information can be distance difference, range difference, etc. In this embodiment, by using the difference information (such as whether the distance exceeds a certain distance value, such as 500m, and this distance value can be determined according to empirical values) between the location coordinates of the geographical elements described by the location POI name and the location coordinates of the location POI, to identify whether the location POI is a wrong POI, so as to obtain the wrong POI information corresponding to the location POI.

[0060] Step S202: Obtain the second wrong POI information of the location POI according to the second difference information between the location coordinates of the second geographical element carried in the address information of the location POI and the location coordinates of the location POI; wherein, the second geographical element is an element obtained by splitting the address of the location POI.

[0061] As mentioned above, when the location POI is published or uploaded, the name and address of the location POI are carried. The address of the location POI is a physical location description of the POI, which can include street, house number, city, postal code, country, etc. Among them, the street, house number, and city correspond to the administrative division description of the location POI address and are the second geographical elements. In addition, the address description may also contain information corresponding to the first geographical element, such as store in XXX Building, XX business district, etc. The first geographical element and the second geographical element are only used to describe similar objects and distinguish the geographical elements carried in the address information and the name information, without other special meanings. The first geographical element and the second geographical element can be the same element or different elements in some scenarios, and this embodiment does not make special limitations on this. The same goes for the first difference information and the second difference information, and no more elaboration will be made here.

[0062] It should be noted that the process of obtaining the location coordinates of the second geographical element and the process of obtaining the second wrong POI information of the location POI are similar to the above-mentioned first geographical element, and the relevant descriptions can refer to the above-mentioned first geographical element, and no more elaboration will be made here.

[0063] Step S203: Determine that the location POI is a wrong POI according to at least one of the first wrong POI information and the second wrong POI information.

[0064] In one example, the location POI can be identified as an incorrect POI by the first geographical element carried in the name, that is, it can be directly determined that the location POI is an incorrect POI, or the location POI can be identified as an incorrect POI by the second geographical element carried in the address, that is, it can be directly determined that the location POI is an incorrect POI. In some examples, when both the first geographical element in the name and the second geographical element in the address identify that the POI is incorrect, it can be determined that the location POI is an incorrect POI. The specific determination method for incorrect POIs in this embodiment is not particularly limited. Or when the name does not identify an incorrect POI, continue to identify through the location POI. If it is identified as an incorrect POI, it is determined that the location POI is an incorrect POI. This embodiment does not make a special limitation on this.

[0065] Through the above technical solution, by splitting the name and address in the location POI, the geographical (address) elements therein are identified, and combined with the difference information between the location coordinates of the location POI, the incorrect data of the mismatch situation existing in the POI address - POI name - POI coordinate triple can be quickly judged. This process can automatically realize the discovery of errors in a large number of POI data without manual participation, reduce the labor cost, and has a higher recognition efficiency, without causing missed recognition. Furthermore, it is convenient to screen out incorrect POIs in the POI library, thus solving the problem that users are difficult to quickly locate the corresponding location point, which affects the user's map usage experience.

[0066] Figure 3 FIG. is a schematic flowchart of another method for processing POI data provided by an embodiment of the present application. On the basis of the above embodiment, this embodiment further exemplifies the acquisition process of the first geographical element. This embodiment uses a name splitting model and an address splitting model to split the name and address of the location POI respectively, which can effectively improve the recognition accuracy of geographical elements. Specifically, in addition to the above steps S201 - S203, this embodiment may further include step S301 and step S302 of splitting the name and address of the location POI.

[0067] It should be noted that in some embodiments, only one of the steps may be adopted. For example, only step S301 or only step S302 is included. For the unexemplified part, other methods can be used to obtain geographical elements (such as the rule matching method or other methods mentioned above). This embodiment does not elaborate too much on this.

[0068] Step S301: According to a preset name splitting model, split the name of the location POI to obtain a name splitting result, where the name splitting result includes at least one first geographical element; wherein, the name splitting model is trained based on historical location POI data and is used to split and identify the first geographical element carried in the name of the location POI.

[0069] In this embodiment, the name segmentation model can be directly trained and used by the server, or can be trained on other devices and transmitted to the server for deployment. This embodiment does not make special limitations on this. During the name segmentation process, the location POI can be input into the trained name segmentation model, and the model is used for processing. The model will segment the POI name into different parts according to the patterns and rules it has learned, and identify the first geographical element therein to obtain the name segmentation result. Optionally, in addition to the identified first geographical element, the name segmentation result may also include the first element type. In the following embodiments, the acquisition process of the first element type may be obtained based on the name segmentation result.

[0070] Exemplarily, the training method of the above name segmentation model may include the following steps: collecting historical POI data, and performing annotation processing on the first geographical elements carried by the names in the historical POI data to obtain a first training sample set; and training a deep learning model for sequence annotation tasks based on the first training sample set to obtain the name segmentation model.

[0071] In this embodiment, the historical POI data can be a set of historical location POI data with a certain amount of data. The geographical elements carried by the names of each data in the large amount of POI data set can be manually annotated to train a model for sequence annotation tasks, such as a BERT+CRF (Conditional Random Field) model, to identify the geographical elements in the POI name, such as (XX Road) branch store / (XX Business District) main point information.

[0072] Optionally, a large amount of historical POI data can be collected. The POI data can include information such as the name, address, and category of the POI. Next, the collected POI data is manually annotated. By annotating the geographical elements carried in the name, the geographical elements can include roads, landmarks, areas of interest (AOIs, such as business districts), etc. During the annotation process, each word or character in the POI name can be marked as a specific geographical element category (such as "road", "landmark") or "non-geographical element". The annotated data is organized into a training sample set, and each sample can consist of the POI name and its corresponding geographical element label. In this embodiment, a deep learning model architecture suitable for sequence annotation tasks can be adopted, such as the BERT+CRF model, which combines the powerful language representation ability of BERT (Bidirectional Encoder Representations from Transformers) and the sequence annotation ability of CRF (Conditional Random Field). In some embodiments, other models such as the model combined with bidirectional long short-term memory network (BiLSTM) and CRF, or other models based on Transformer, etc. can also be adopted. The selected model is trained using the first training sample set. During the training process, the model can learn how to predict the corresponding geographical element label according to the input POI name sequence based on a large amount of data sets, and when the loss condition or iteration condition is met, a trained model is obtained.

[0073] It can be understood that the sequence annotation task is a basic task in natural language processing (NLP), which can assign a label to each element in the input sequence. In this embodiment, a deep learning model for the sequence annotation task is trained to obtain a name segmentation model, which is convenient for efficiently identifying and annotating geographical elements in the name and can output the corresponding element types.

[0074] Step S302: According to the preset address segmentation model, perform address segmentation and recognition on the location POI to obtain an address segmentation result, where the address segmentation result includes at least one second geographical element; wherein, the address segmentation model is trained based on historical location POI data and is used to segment and recognize the second geographical elements carried in the address of the location POI.

[0075] It should be noted that the process of obtaining the second geographical element is similar to the process of obtaining the first geographical element described above, and related descriptions will not be elaborated here.

[0076] Exemplarily, the training method of the above address segmentation model may be: collecting historical POI data, performing annotation processing on the second geographical elements carried in the addresses in the historical POI data to obtain a second training sample set, and training a deep learning model for sequence annotation tasks to obtain the address segmentation model.

[0077] It should be noted that the training process of the address segmentation model is similar to the training process of the above name segmentation model, and relevant descriptions will not be elaborated here.

[0078] Through the above technical solution, by training the name segmentation model and the address segmentation model, the names and addresses in the location POI are segmented and recognized to obtain the corresponding first geographical elements and second geographical elements, which can effectively improve the recognition accuracy of geographical elements and further improve the recognition accuracy of incorrect POIs.

[0079] Figure 4 It is a flowchart of another method for processing POI data provided by an embodiment of the present application. On the basis of the above embodiment, this embodiment identifies incorrect POIs according to different element types to further improve the recognition accuracy. Specifically, in addition to the above steps S201 - S203, the method provided in this embodiment may further include step S401, and step S201 is further divided into step S2011.

[0080] Step S401: Obtain the first element type of the first geographical element according to the preset name segmentation model, where the first element type includes at least one of the following: road element, landmark element, and geographical interest area element.

[0081] Exemplarily, after the name segmentation model performs name segmentation processing on the location POI, it can output the segmented and recognized geographical elements and the corresponding element types. In other examples, the element types of geographical elements can also be recognized according to other rules, such as recognizing element types according to mapping relationships, etc.

[0082] Step S2011: Obtain the first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographical element carried in the name of the location of interest POI and the position coordinates of the location POI.

[0083] In this embodiment, by combining the corresponding element types, the differences between the position coordinates of geographical elements and the POI coordinates are respectively recognized, such as distance, crossing roads, whether it is within the business district range, etc., and the corresponding difference information is combined to determine whether the location POI is an incorrect POI, so as to accurately recognize various geographical elements.

[0084] In an alternative embodiment, the first difference information includes a distance difference. The above step S2021 may obtain the first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographical element and the position coordinates of the location POI, and may be the following steps:

[0085] If the first geographical element is a road element, obtain the distance difference between the position coordinates of the road element and the position coordinates of the location POI. When the distance difference reaches a first preset distance threshold, obtain the first incorrect POI information of the location POI regarding the road distance.

[0086] Among them, a road element, that is, an element carrying XXX Road or XXX Street (Lane) in the location POI name. By calculating the distance difference between the actual position coordinates of the XXX Road and the position coordinates of the location POI, it is determined whether the location POI is an incorrect POI. Exemplarily, a geographical calculation method (such as the Haversine formula) can be used to calculate the distance between the POI coordinates and the coordinates of "XXX Road", or for a linear geographical object such as a road, the distance from the POI to the nearest point on the road can also be calculated, and the distance difference can be calculated. When the distance difference reaches the first preset distance threshold (those skilled in the art can adaptively set the first preset distance threshold in combination with actual applications or empirical values, such as 500 - 1000m, which can be adjusted according to different cities and road conditions), it is considered that the location POI is determined as an incorrect POI during the name recognition process, and the corresponding first incorrect POI information is obtained. The first incorrect POI information indicates that the name recognition process determines it as an incorrect POI.

[0087] In another alternative embodiment, the first difference information may also be cross-road information. Specifically, the above step S2021 may be the following steps: If the first geographical element is a road element, obtain the cross-road information between the position coordinates of the road element and the position coordinates of the location POI. The cross-road information is used to indicate whether there are other roads except the road element in the straight-line area connecting the position coordinates of the road element and the position coordinates of the location POI. If so, obtain the first incorrect POI information of the location POI regarding the cross-road.

[0088] In this embodiment, the cross-road information can be determined by this device (i.e., the server) and obtained from the memory (or other components) during use, or it can be obtained from other devices or servers. Exemplarily, for the determination of cross-road information, a straight line between the position coordinates of the road element and the position coordinates of the POI can be determined. This straight line represents the most direct path from the road to the POI. In this straight line area, it can be checked whether there are other roads, which can be determined by obtaining the road network information around the position coordinates of the road element, such as through a GIS system or a map API, etc., to detect all road data within the straight line area. If other roads other than the said road element are detected in the straight line area, it indicates that there is a cross-road phenomenon between the POI and its associated road.

[0089] In some alternative embodiments, the distance difference and the cross-road information can also be combined to jointly identify whether a POI is determined to be an incorrect POI when the road element is carried in the name of the location POI. Exemplarily, as Figure 5 shown, although the distance from the starting point to Street A is 300m, which is less than the first preset threshold, according to the road network information corresponding to the position coordinates of Street A, it can be queried that there are Road B and Road C between the POI and the road. In this case, it is considered that there is a cross-road. If the address of the POI at this location is No. xx, Street A, it is still considered to be incorrect POI data.

[0090] Through the above methods of calculating the distance difference and identifying the cross-road, it is possible to quickly determine whether a POI is an incorrect POI or potential incorrect POI data by the name of the location POI, and it is possible to identify the POIs that may be mislabeled due to road interference, thereby improving the accuracy and reliability of the POI data.

[0091] In some embodiments, when the geographical element is a landmark element, it is also possible to determine whether it is an incorrect POI according to the distance difference. Specifically, the above step S2021 can obtain the first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographical element and the position coordinates of the location POI, as follows: if the first geographical element is a landmark element, obtain the distance difference between the position coordinates of the landmark element and the position coordinates of the location POI. When the distance difference reaches the second preset distance threshold, obtain the first incorrect POI information of the location POI regarding the landmark distance; wherein, the second preset distance threshold is less than or equal to the first preset distance threshold.

[0092] In this embodiment, calculating the distance difference between the position coordinates of the landmark element and the position coordinates of the position POI is similar to the way of calculating the distance difference of the above road elements, and the relevant description will not be elaborated too much. When identifying whether the landmark distance is incorrect POI information, it is determined that the distance threshold for the distance difference is smaller or equal to the threshold distance for the distance difference of the road element, that is, the second preset distance threshold is less than or equal to the first preset distance threshold. This is because it is considered that landmarks are usually elements of point-like or small-scale areas with significant features, and users have a higher expectation for their accuracy in actual navigation. That is, landmarks are usually used for precise positioning or as navigation reference points. Therefore, higher accuracy is required to ensure that users can accurately identify and reach the landmark. And roads are linear elements that usually cover a large area. Due to the length and width of the road itself, when users use the road for navigation, a larger error range may be allowed, and a larger distance error may not significantly affect the navigation effect.

[0093] For the element obtained in the name that is a landmark POI (eg: Cut out the landmark: XXX Building from C Store (XXX Building)), the relationship between the coordinates of the landmark and the POI coordinates is directly calculated, that is, if the distance between the coordinates of C Store (XXX Building) and XXX Building (assuming multiple XXX Buildings are recognized, it is the XXX Building closest to C Store) exceeds the second preset threshold (assuming the first preset threshold is 1000m, the second preset threshold can be 500m), it is considered incorrect data.

[0094] It can be understood that a landmark is a location or iconic building with significant features or historical significance within a certain area. It can be naturally formed (such as mountains, lakes) or man-made (such as towers, monuments, bridges), etc. For example, in the above example, XXX Building is the recognized landmark carried in the name with C Store (XXX Building).

[0095] Through the above method of identifying landmark distances, POI data with incorrect landmark information carried in the name can be effectively identified. And for these two different types of geographical elements, landmarks and roads, setting different distance thresholds can better meet the user's high-precision requirements for landmarks, while allowing a larger error range for roads, which is more in line with actual applications and brings a better experience to users.

[0096] In some embodiments, the type of geographical element may also be a geographical area of interest (AOI), and the first difference information may also be cross-area-of-interest information. The above step S2021 of obtaining the first incorrect POI information of the position POI according to the first element type and the first difference information between the position coordinates of the first geographical element and the position coordinates of the position POI may include the following steps: If the first geographical element is a geographical area-of-interest element, obtain the cross-area-of-interest information between the position coordinates of the geographical area-of-interest element and the position coordinates of the position POI. The cross-area-of-interest information is used to indicate whether the position coordinates of the position POI are within the position coordinates of the geographical area-of-interest element. If not, obtain the first incorrect POI information of the position POI regarding the cross area of interest.

[0097] For the element obtained in the name which is a geographical area of interest (AOI), an AOI is a special type of POI. It is not a point on the map but a large POI with an area, such as a park, a scenic area, a business district, and so on. Exemplarily, for a C-store (XXX business district), XXX business district can be replaced with any business district name known to the user. The XXX business district is an AOI. As Figure 6 shown, AOIs are some large and important POIs that have been verified by special personnel during the map-making process and have a relatively high accuracy rate. If the coordinates are located outside the AOI (the corresponding position coordinates, i.e., the boundary coordinates), it is considered incorrect POI data.

[0098] Exemplarily, the process of obtaining cross-area-of-interest information can be as follows: Use the geocoding system or the GIS system to obtain the position coordinates of the geographical area-of-interest element (AOI). These position coordinates are the boundary coordinates of the area, and these boundary coordinates define the geographical scope of the AOI. By calculating whether the position coordinates of the POI are within the boundary coordinates of the AOI, for example, through geospatial analysis, use the algorithm of point-in-polygon to determine whether the position coordinates of the POI are within the coordinate range of the AOI, that is, within the boundary coordinates. If the position coordinates of the POI are within the boundary of the AOI, it indicates that the geographical location of the POI is consistent with the AOI information carried in its name. Otherwise, it indicates that there is a cross-area-of-interest situation, that is, the position of the POI is inconsistent with the AOI indicated in its name, and this position POI is a potential incorrect POI. In this way, the incorrect information in the POI when it carries AOI geographical elements can be quickly identified.

[0099] It should be noted that in addition to the above-mentioned road elements, landmark elements, and geographical area-of-interest elements, the road elements may also include other road elements, for example, public transportation elements (such as XXX station), etc. This embodiment does not make special limitations on this.

[0100] Correspondingly, this embodiment further exemplifies the recognition process of geographical elements carried in the POI address, so as to identify potential incorrect POIs with incorrect addresses, thereby further improving the screening accuracy of incorrect POIs. Specifically, the method provided in this embodiment may further include the following steps: obtaining the second element type of the second geographical element according to a preset address segmentation model, where the second element type includes at least one of administrative division elements and first element type elements.

[0101] It should be noted that the process of obtaining the second geographical element in this embodiment is similar to the above-mentioned process of obtaining the first geographical element, and related descriptions will not be elaborated here. Among them, administrative division elements are relatively common elements in the geographical elements carried in the address (in some embodiments, the administrative division element type may also be carried in the POI name, and this embodiment does not make special limitations on this), such as XXX County, XXX City, XXX Province, XXXX Town.

[0102] The above step S202 of obtaining the second incorrect POI information of the location POI according to the second difference information between the position coordinates of the second geographical element carried in the address information of the location POI and the position coordinates of the location POI may include the following steps:

[0103] If the second geographical element is an administrative division element, it is successively determined whether the position coordinates of the location POI are within the range of each level of administrative division corresponding to the position coordinates of the administrative division element. If not, the second incorrect POI information of the location POI regarding the administrative division is obtained; and / or,

[0104] If the second geographical element is a first element type element, the second incorrect POI information of the location POI regarding at least one of roads, landmarks, and cross-interest areas is obtained according to the corresponding difference information between the second geographical element and the location POI.

[0105] Exemplarily, the district / county, township, village, road, road + house number, and landmark POI information segmented from the address can be obtained, and the geocoding service can be called for the address to successively determine whether the POI coordinates are within the ranges of the corresponding district / county, township, and village. If one of them is not, it is determined as incorrect POI data; if there is information such as roads / AOIs / landmarks in the address, it is determined whether the position of the source POI is near the road / AOI / landmark, and the judgment method is the same as that of the above first geographical element. In some examples, the geocoding service can also be called for the address, and if the position coordinates corresponding to the address returned and the position coordinates of the POI differ by more than a certain distance (such as 2 km), it is considered an incorrect POI.

[0106] As mentioned above, in this embodiment, the second geographic element can be the same element as the first geographic element. When they are the same type of elements, the method of identifying the wrong POI is the same as when the name carries the first geographic element. For example, the second geographic element carries a road element. By calculating the distance difference between the position coordinates of the road element and the position coordinates of the position POI or the cross-road information, it is determined whether it is a potential wrong POI with wrong address information. Landmarks, area of interest elements, etc. are similar and will not be elaborated here.

[0107] In an exemplary embodiment, the POI coordinates, name, and address are combined to perform the wrong POI identification process, which can be as follows: Figure 7 As shown, the following process is included: a. In the initialization state, input the location POI to be identified (it can be one or a batch), and the location POI includes a name, an address and coordinates; b. Use the name segmentation model to segment the location POI to determine whether there is a branch name / main point information. The main point information is usually described using the first geographical element, that is, a road / landmark / AOI, etc., if the main point information exists; execute c. Search the location coordinates of the corresponding main point through the geocoding service to determine whether the main point is near the location POI (determine whether it is near according to the specific element type of the main point, such as within 1000m for road type and within 500m for landmark type); if it is not near, obtain the first error POI information determined by the corresponding POI name (that is, the name is determined to be a potential problem POI), and the location POI can be directly determined to be an error POI. If it is nearby, it means that the POI name is determined to be successful (or the name is determined to have no main point information, that is, the first geographical element does not exist), and step d is executed to call the geocoding service for the POI address to obtain the address coordinates, and step e is executed to determine whether the address coordinates and the POI location coordinates exceed a certain distance (such as 2km). If so, it is considered to be a potential problem POI (corresponding to the second error POI information). In this example, it can be directly identified as an error POI; if it does not exceed the distance, step f is executed to segment and identify the address of the location POI using the address segmentation model, and obtain the second geographical elements such as township, county, road, landmark, AOI, etc. in the address, and step g is executed to determine the difference between the POI coordinates and each element in turn. If the difference between the coordinates and the elements is too large (such as the distance difference reaches the corresponding distance threshold, is not within the corresponding administrative division, etc.), the location POI is considered to be an error POI.

[0108] It can be understood that the above examples are only possible examples of the embodiments of the present application. In some examples, when it is determined as a potential error POI, other determination methods can be further combined to further identify whether the POI at this location is an error POI. For example, when it is identified as an error POI using the name segmentation result, the address segmentation result can be further combined to determine whether it is also identified as an error POI. When both are error POIs, it is considered that the POI at this location is an error POI; alternatively, the POI address can be segmented and identified first, and then the POI name can be segmented and identified. This embodiment does not make special limitations on this.

[0109] Figure 8 This is a schematic flowchart of another method for processing POI data provided by the embodiments of the present application. On the basis of the above embodiments, this embodiment further exemplifies the screening of error POIs after identifying error POIs, and in addition to the above method of screening error POIs, the data in the library is also clustered as a whole in combination with the geographical elements in the address to screen out outlier POIs, thereby further improving the efficiency and accuracy of screening error POIs. Specifically, the location POI in this embodiment is any POI in the POI database, such as Figure 8 As shown, in addition to the above steps S201 - S203, the following steps may further be included:

[0110] S801. Screen out all location POIs determined to be error POIs from the POI database.

[0111] In one example, the server can construct an SQL statement for deleting the determined error POIs and use this SQL statement to screen out the corresponding location POIs, which is convenient for batch screening of error POIs. In some examples, the method of this embodiment can also be applied to the POI database. After identifying the error POIs, they can be directly screened out in the POI database.

[0112] Furthermore, in combination with Figure 9a and Figure 9b As shown, the method of clustering is used to screen out abnormal outliers to improve the accuracy of identifying error POIs. Specifically, the following steps may further be included:

[0113] According to the address segmentation model, obtain the address segmentation results of all location POIs in the POI database to obtain the second geographical elements of each location POI;

[0114] Cluster the location POIs with the same second geographical elements to obtain a clustering result;

[0115] For the clustered location POIs in the clustering result, calculate the location distances between the clustered location POIs, and screen out the error POIs with location distances greater than the third preset threshold from the clustered location POIs according to the location distance results.

[0116] In this embodiment, all addresses in the POI library are segmented to screen out those from which a road + house number or landmark building can be segmented (landmark buildings are a series of representative POIs manually selected in a city, such as parks, scenic spots, business districts, etc., such as XX Park, XXX Business District, XX Museum). The coordinates of POIs with the same road + house number or landmark building segmented from the address are clustered, and outliers are found. As Figure 9b shown, taking the clustering of POIs with addresses all at No. 10 XX Road as an example, and labeling each location POI with a number. Among them, the POI labeled 5 is far from other location POIs and is an outlier. The location POI corresponding to the label 5 is determined as incorrect POI data and is screened out.

[0117] Exemplarily, the address element segmentation system is used to segment the road house number or landmark building from the POI addresses in the library. POIs with the same road house number / landmark building are considered to belong to the same cluster. The outlier calculation method can be that the physical distances between all POIs in the same class are calculated pairwise, and a POI whose distance from other POIs in the class exceeds 2 km (in some embodiments, other distances can also be determined according to empirical values, and this embodiment does not limit this) is considered an outlier. Through the above technical solution, that is, it is not an incorrect POI. In this way, the method of clustering data in the library by combining geographical elements screens out outliers by clustering the massive data in the library according to the geographical elements carried in the address, so as to screen out the corresponding incorrect POIs, further improving the screening efficiency and accuracy of incorrect POIs. Especially when there is a large amount of incorrect POI data, the screening efficiency is more obvious.

[0118] Figure 10 is a schematic structural diagram of a point of interest data processing device provided by an embodiment of the present application. As Figure 10 shown, the device 1000 includes a first recognition module 1001, a second recognition module 1002, and an incorrect POI determination module 1003, where

[0119] The first recognition module 1001 is configured to obtain first incorrect POI information of the location POI according to the first difference information between the position coordinates of the first geographical element carried in the name of the location point of interest POI and the position coordinates of the location POI; where the first geographical element is an element obtained by segmenting the name of the location POI.

[0120] A second recognition module 1002, configured to obtain second incorrect POI information of the location POI according to second difference information between the position coordinates carried in the address information of the location POI and the position coordinates of the location POI; wherein, the second geographical element is an element obtained by splitting the address of the location POI.

[0121] An incorrect POI determination module 1003, configured to determine that the location POI is an incorrect POI according to at least one of the first incorrect POI information and the second incorrect POI information.

[0122] In one implementation, the device further includes a name splitting module for splitting the name of the location POI, configured to split the name of the location POI according to a preset name splitting model to obtain a name splitting result, where the name splitting result includes at least one of the first geographical elements; wherein, the name splitting model is trained based on historical location POI data and is used to split and identify the first geographical elements carried in the name of the location POI.

[0123] In one implementation, the device further includes an address splitting module for splitting the address of the location POI, configured to perform address splitting and identification on the location POI according to a preset address splitting model to obtain an address splitting result, where the address splitting result includes at least one of the second geographical elements; wherein, the address splitting model is trained based on historical location POI data and is used to split and identify the second geographical elements carried in the address of the location POI.

[0124] In one implementation, the device further includes:

[0125] A first element type obtaining module, configured to obtain a first element type of the first geographical element according to a preset name splitting model, where the first element type includes at least one of the following: road element, landmark element, and geographical interest area element.

[0126] The first recognition module 1001 is specifically configured to obtain first incorrect POI information of the location POI according to the first element type and first difference information between the position coordinates of the first geographical element and the position coordinates of the location POI.

[0127] In one implementation, the first difference information includes a distance difference and / or cross-road information.

[0128] The first recognition module 1001 is specifically configured as follows: If the first geographical element is a road element, obtain the distance difference between the position coordinates of the road element and the position coordinates of the position POI. When the distance difference reaches a first preset distance threshold, obtain the first incorrect POI information of the position POI regarding the road distance; and / or, obtain the cross-road information between the position coordinates of the road element and the position coordinates of the position POI, where the cross-road information is used to indicate whether there are other roads except the road element in the straight-line area connecting the position coordinates of the road element and the position coordinates of the position POI. If so, obtain the first incorrect POI information of the position POI regarding crossing the road.

[0129] In one implementation, the first difference information may include a distance difference. The first recognition module 1001 is further configured as follows: If the first geographical element is a landmark element, obtain the distance difference between the position coordinates of the landmark element and the position coordinates of the position POI. When the distance difference reaches a second preset distance threshold, obtain the first incorrect POI information of the position POI regarding the landmark distance; where the second preset distance threshold is less than or equal to the first preset distance threshold.

[0130] In one implementation, the first difference information includes cross-interest area information; the first recognition module 1001 is specifically configured as follows: If the first geographical element is a geographical interest area element, obtain the cross-interest area information between the position coordinates of the geographical interest area element and the position coordinates of the position POI, where the cross-interest area information is used to indicate whether the position coordinates of the position POI are within the position coordinates of the geographical interest area element. If not, obtain the first incorrect POI information of the position POI regarding crossing the interest area.

[0131] In one implementation, it further includes a first training module for training the name segmentation model, which is configured to collect historical POI data, perform annotation processing on the first geographical features carried by the names in the historical POI data to obtain a first training sample set; and, based on the first training sample set, train a deep learning model for sequence labeling tasks to obtain the name segmentation model;

[0132] In one implementation, it further includes a second training module for training the address segmentation model, which is configured to collect historical POI data, perform annotation processing on the second geographical features carried by the addresses in the historical POI data to obtain a second training sample set, and train a deep learning model for sequence labeling tasks to obtain the address segmentation model.

[0133] In one implementation, the device further includes:

[0134] A second element type acquisition module, configured to acquire a second element type of the second geographical element according to a preset address segmentation model, where the second element type includes at least one of an administrative division element and a first element type element;

[0135] The second recognition module 1002 is specifically configured to: if the second geographical element is an administrative division element, sequentially determine whether the location coordinates of the location POI are within the administrative division ranges corresponding to the location coordinates of the administrative division element. If not, acquire second incorrect POI information of the location POI regarding the administrative division; and / or, if the second geographical element is a first element type element, acquire second incorrect POI information of the location POI regarding at least one of a road, a landmark, and an area across interests according to the corresponding difference information between the second geographical element and the location POI.

[0136] In one implementation, the location POI is any POI in the POI database; the apparatus further includes: an incorrect POI screening module, configured to screen out all location POIs determined to be incorrect POIs from the POI database.

[0137] In one implementation, the apparatus further includes:

[0138] A segmentation acquisition module, configured to acquire an address segmentation result of all location POIs in the POI database according to an address segmentation model, so as to obtain a second geographical element for each location POI;

[0139] A clustering module, configured to cluster location POIs having the same second geographical element to obtain a clustering result;

[0140] The incorrect POI screening module is further configured to calculate a location distance between each clustering location POI for the clustering location POIs in the clustering result, and screen out incorrect POIs with a location distance greater than a third preset threshold from the clustering location POIs according to the location distance result.

[0141] For relevant descriptions, reference may be made to the corresponding descriptions and effects of the steps in the method embodiment of the present application, and details are not described herein again.

[0142] Figure 11 The structure diagram of a server provided by an embodiment of the present application is shown as Figure 11 shown. The server 1100 includes: a processor 1101, and a memory 1102 communicatively connected to the processor 1101;

[0143] The memory 1102 stores computer-executable instructions;

[0144] The processor 1101 executes the computer-executable instructions stored in the memory 1102 to implement the method for processing point-of-interest data of the high-precision map. The memory 1102 and the processor 1101 are connected through a bus 1103.

[0145] For relevant descriptions, reference can be made to the corresponding descriptions and effects of the steps in the method embodiments of this application, and details will not be elaborated here.

[0146] The embodiments of the present application further provide a computer-readable storage medium storing computer-executable instructions, which are used to implement the method for processing point-of-interest data provided in the above method embodiments when executed by a processor.

[0147] The computer-readable storage medium may be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0148] For relevant descriptions, reference can be made to the corresponding descriptions and effects of the steps in the method embodiments of this application, and details will not be elaborated here.

[0149] The embodiments of the present application further provide a computer program product, which includes a computer program that implements the method for processing point-of-interest data provided in the above method embodiments when executed by a processor.

[0150] For relevant descriptions, reference can be made to the corresponding descriptions and effects of the steps in the method embodiments of this application, and details will not be elaborated here.

[0151] The embodiments of the present application further provide a chip, including a memory and a processor. The memory is used to store a computer program, and the processor is used to call and run the computer program from the memory to execute the method for processing point-of-interest data.

[0152] For relevant descriptions, reference can be made to the corresponding descriptions and effects of the steps in the method embodiments of this application, and details will not be elaborated here.

[0153] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of devices or modules can be in electrical, mechanical or other forms.

[0154] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the application disclosed herein. The present application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include known common knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and embodiments are only regarded as exemplary, and the true scope and spirit of the present application are pointed out by the following claims.

[0155] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope. The scope of the present application is only limited by the appended claims.

Claims

1. A method for processing point-of-interest data, characterized in that, Including: Obtaining first incorrect POI information of the location POI according to first difference information between position coordinates of a first geographic element carried in a name of the location point of interest (POI) and the position coordinates of the location POI; wherein, the first geographic element is an element obtained by splitting the name of the location POI; Obtaining second incorrect POI information of the location POI according to second difference information between position coordinates of a second geographic element carried in address information of the location POI and the position coordinates of the location POI; wherein, the second geographic element is an element obtained by splitting the address of the location POI; Determining that the location POI is an incorrect POI according to at least one of the first incorrect POI information and the second incorrect POI information.

2. The method according to claim 1, wherein: The method further includes: obtaining a first element type of the first geographic element according to a preset name splitting model, and the first element type includes at least one of the following: road element, landmark element, and geographic interest area element; The obtaining first incorrect POI information of the location POI according to first difference information between position coordinates of a first geographic element carried in a name of the location point of interest (POI) and the position coordinates of the location POI includes: Obtaining first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographic element and the position coordinates of the location POI.

3. The method according to claim 2, wherein The first difference information includes a distance difference value and / or cross-road information; The obtaining first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographic element and the position coordinates of the location POI includes: If the first geographic element is a road element, obtaining a distance difference value between the position coordinates of the road element and the position coordinates of the location POI, and when the distance difference value reaches a first preset distance threshold, obtaining first incorrect POI information of the location POI regarding road distance; and / or, Obtaining cross-road information between the position coordinates of the road element and the position coordinates of the location POI, where the cross-road information is used to indicate whether there is any other road except the road element in a straight-line area connecting the position coordinates of the road element and the position coordinates of the location POI, and if so, obtaining first incorrect POI information of the location POI regarding cross-road.

4. The method according to claim 2, wherein The first difference information includes a distance difference value; The obtaining first incorrect POI information of the location POI according to the first element type and the first difference information between the position coordinates of the first geographic element and the position coordinates of the location POI includes: If the first geographical element is a landmark element, obtain the distance difference between the position coordinates of the landmark element and the position coordinates of the position POI. When the distance difference reaches a second preset distance threshold, obtain the first incorrect POI information of the position POI regarding the landmark distance; Wherein, the second preset distance threshold is less than or equal to the first preset distance threshold.

5. The method according to claim 2, wherein The first difference information includes cross - interest - area information; The obtaining of the first incorrect POI information of the position POI according to the first element type and the first difference information between the position coordinates of the first geographical element and the position coordinates of the position POI includes: If the first geographical element is a geographical interest - area element, obtain the cross - interest - area information between the position coordinates of the geographical interest - area element and the position coordinates of the position POI. The cross - interest - area information is used to indicate whether the position coordinates of the position POI are within the position coordinates of the geographical interest - area element. If not, obtain the first incorrect POI information of the position POI regarding the cross - interest - area.

6. The method according to any one of claims 1 - 5, characterized in that, The method further includes: obtaining the second element type of the second geographical element according to a preset address segmentation model, where the second element type includes at least one of administrative division elements and first element type elements; The obtaining of the second incorrect POI information of the position POI according to the second difference information between the position coordinates of the second geographical element carried in the address information of the position POI and the position coordinates of the position POI includes: If the second geographical element is an administrative division element, sequentially determine whether the position coordinates of the position POI are within the administrative division ranges corresponding to the position coordinates of the administrative division element. If not, obtain the second incorrect POI information of the position POI regarding the administrative division; and / or, If the second geographical element is a first element type element, obtain the second incorrect POI information of the position POI regarding at least one of roads, landmarks, and cross - interest - areas according to the corresponding difference information between the second geographical element and the position POI.

7. The method according to any one of claims 1 - 5, characterized in that, The segmenting of the name of the position POI includes: Segmenting the name of the position POI according to a preset name segmentation model to obtain a name segmentation result, where the name segmentation result includes at least one of the first geographical elements; wherein, the name segmentation model is trained based on historical position POI data and is used to segment and identify the first geographical elements carried in the name of the position POI; and / or, The segmenting of the address of the position POI includes: According to a preset address segmentation model, perform address segmentation recognition on the location POI to obtain an address segmentation result, where the address segmentation result includes at least one of the second geographical elements; wherein, the address segmentation model is trained based on historical location POI data and is used to segment and recognize the second geographical elements carried in the address of the location POI.

8. The method according to any one of claims 1-5, characterized in that It further includes: According to the address segmentation model, obtain the address segmentation results of all location POIs in the POI database to obtain the respective second geographical elements of each location POI; Cluster the location POIs with the same second geographical element to obtain a clustering result; For the clustered location POIs in the clustering result, calculate the location distances between the respective clustered location POIs, and filter out the incorrect POIs with location distances greater than a third preset threshold from the clustered location POIs according to the location distance results.

9. An interest point data processing device, characterized in that, It includes: A first recognition module configured to obtain first incorrect POI information of the location POI according to the first difference information between the position coordinates of the first geographical element carried in the name of the location of interest POI and the position coordinates of the location POI; wherein, the first geographical element is an element obtained by segmenting the name of the location POI. A second recognition module configured to obtain second incorrect POI information of the location POI according to the second difference information between the position coordinates of the second geographical element carried in the address information of the location POI and the position coordinates of the location POI; wherein, the second geographical element is an element obtained by segmenting the address of the location POI. An incorrect POI determination module configured to determine the location POI as an incorrect POI according to at least one of the first incorrect POI information and the second incorrect POI information.

10. A server / computer-readable storage medium / computer program product, characterized in that, The server includes: a memory and a processor; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory, so that the server executes the POI data processing method according to any one of claims 1 to 8; and / or, The computer-readable storage medium stores computer execution instructions, and when the computer execution instructions are executed by the processor, they are used to implement the POI data processing method according to any one of claims 1-8; and / or, The computer program product includes a computer program, and when the computer program is executed by the processor, it implements the POI data processing method according to any one of claims 1-8.

Images