Geographic information retrieval methods, devices, equipment, and media based on large language models

Abstract

This disclosure presents embodiments of a geographic information retrieval method, apparatus, device, and medium based on a large language model. One specific implementation of the method includes: acquiring target geographic text information input by a target user terminal; performing recognition processing on the target geographic text information based on a semantic parser to generate a text feature information set; acquiring preset generation rules and generating a spatial operation sequence based on the preset generation rules; performing retrieval preprocessing on the spatial operation sequence based on the spatial operation sequence to generate a preprocessed spatial operation sequence; performing stage retrieval processing based on the preprocessed spatial operation sequence to generate geographic entity information; generating a geographic entity page corresponding to the geographic entity information; and sending the geographic entity page to the target user terminal for display. This implementation accurately generates user query results and avoids wasting transmission resources.

Description

Technical Field

[0001] Embodiments of this disclosure relate to the field of computer technology, and more specifically to geographic information retrieval methods, apparatus, devices, and media based on large language models. Background Technology

[0002] When users perform location searches, they typically input rather vague phrases. Currently, the common approach for location searches using vague terms is to perform fuzzy queries or similarity searches against a pre-defined thesaurus, and then sort the results according to their probability.

[0003] However, when using the above method for geolocation retrieval, the following technical problems often arise:

[0004] The inability to process ambiguous user input results in a large number of results that do not match the user's search purpose, leading to a waste of transmission resources when transmitting the results to the user.

[0005] The information disclosed in this background section is only intended to enhance the understanding of the background of the inventive concept, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The summary portion of this disclosure is intended to provide a brief overview of the concepts, which will be described in detail in the detailed description portion. This summary portion is not intended to identify key or essential features of the claimed technical solutions, nor is it intended to limit the scope of the claimed technical solutions.

[0007] Some embodiments of this disclosure propose geographic information retrieval methods, apparatuses, devices, and computer-readable media based on large language models to address one or more of the technical problems mentioned in the background section above.

[0008] In a first aspect, some embodiments of this disclosure provide a geographic information retrieval method based on a large language model. The method includes: acquiring target geographic text information input by a target user terminal; performing recognition processing on the target geographic text information based on a semantic parser to generate a text feature information set, wherein the text feature information in the text feature information set is one of the following: spatial relationship entities, geographic attribute information, and logical relationship information; acquiring preset generation rules, and generating a spatial operation sequence based on the preset generation rules; performing retrieval preprocessing on the spatial operation sequence based on the spatial operation sequence to generate a preprocessed spatial operation sequence; performing stage retrieval processing based on the preprocessed spatial operation sequence to generate geographic entity information; generating a geographic entity page corresponding to the geographic entity information; and sending the geographic entity page to the target user terminal for display.

[0009] Secondly, some embodiments of this disclosure provide a geographic information retrieval device based on a large language model. The device includes: a first acquisition unit configured to acquire target geographic text information input by a target user terminal; an identification unit configured to perform identification processing on the target geographic text information based on a semantic parser to generate a text feature information set, wherein the text feature information in the text feature information set is one of the following: spatial relationship entities, geographic attribute information, and logical relationship information; a second acquisition unit configured to acquire preset generation rules and generate a spatial operation sequence based on the preset generation rules; a retrieval preprocessing unit configured to perform retrieval preprocessing on the spatial operation sequence based on the spatial operation sequence to generate a preprocessed spatial operation sequence; a stage retrieval unit configured to perform stage retrieval processing based on the preprocessed spatial operation sequence to generate geographic entity information; and a generation unit configured to generate a geographic entity page corresponding to the geographic entity information and send the geographic entity page to the target user terminal for display.

[0010] Thirdly, some embodiments of this disclosure provide an electronic device, including: one or more processors; and a storage device having one or more programs stored thereon, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the method described in any implementation of the first aspect above.

[0011] Fourthly, some embodiments of this disclosure provide a computer-readable medium having a computer program stored thereon, wherein the program, when executed by a processor, implements the method described in any of the implementations of the first aspect above.

[0012] The above embodiments of this disclosure have the following beneficial effects: the geographic information retrieval method based on a large language model in some embodiments of this disclosure avoids the waste of transmission resources. Specifically, the waste of transmission resources is caused by the inability to process fuzzy statements input by users, resulting in a large number of results that do not match the user's search purpose, which leads to the need to consume a lot of transmission resources when transmitting the results to the user, thus wasting transmission resources. Based on this, the geographic information retrieval method based on a large language model in some embodiments of this disclosure first obtains the target geographic text information input by the target user terminal. Thus, the query information input by the user can be determined. Second, based on a semantic parser, the above target geographic text information is identified and processed to generate a text feature information set. Thus, the text features can be parsed by the semantic parser. Then, a preset generation rule is obtained, and a spatial operation sequence is generated based on the above preset generation rule. Thus, the retrieval order can be determined according to the preset rule. Afterward, based on the above spatial operation sequence, the spatial operation sequence is preprocessed to generate a preprocessed spatial operation sequence. Thus, retrieval preprocessing can be performed before retrieval, thereby concretizing fuzzy statements. Next, based on the preprocessed spatial operation sequence described above, a stage retrieval process is performed to generate geographic entity information. Thus, a retrieval can be performed using specific query information, generating geographic entity information that matches the user's query. Finally, a geographic entity page corresponding to the aforementioned geographic entity information is generated and sent to the target user terminal for display. This ensures accurate search results are sent to the user, avoiding the generation of numerous useless search results and thus preventing the waste of transmission resources. Attached Figure Description

[0013] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and elements are not necessarily drawn to scale.

[0014] Figure 1 This is a flowchart of some embodiments of the geographic information retrieval method based on a large language model according to the present disclosure;

[0015] Figure 2 These are schematic diagrams illustrating the structure of some embodiments of the geographic information retrieval device based on a large language model according to the present disclosure;

[0016] Figure 3 This is a schematic diagram of the structure of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Implementation

[0017] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0018] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0019] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0020] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0021] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0022] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0023] Figure 1 A flow 100 of some embodiments of the geographic information retrieval method based on a large language model according to the present disclosure is shown. This geographic information retrieval method based on a large language model includes the following steps:

[0024] Step 101: Obtain the target geographic text information input by the target user terminal.

[0025] In some embodiments, the executing entity (e.g., a server) of the geographic information retrieval method based on a large language model can obtain target geographic text information input by a target user terminal. The target user terminal can be a user terminal with retrieval permissions connected to the executing entity via a wired or wireless connection. The target geographic text information can be the query text input by the user for retrieving geographic entities.

[0026] Step 102: Based on the semantic parser, the target geographic text information is identified and processed to generate a text feature information set.

[0027] In some embodiments, the aforementioned execution entity may perform recognition processing on the aforementioned target geographic text information based on a semantic parser to generate a text feature information set. Here, the aforementioned semantic parser may be a large language model with semantic parsing capabilities. The text feature information in the aforementioned text feature information set may be one of the following: spatial relationship entities, geographic attribute information, and logical relationship information.

[0028] In practice, the above-mentioned target geographic text information can be identified and processed through the following steps:

[0029] The first step is to perform geographic term pre-segmentation on the target geographic text information to generate pre-segmented geographic terms, resulting in a pre-segmented geographic term set. For example, "Chaoyang District of Beijing and north of the Yellow River" can be pre-segmented into "Beijing / Chaoyang District / and / north of the Yellow River".

[0030] The second step involves performing a secondary segmentation process on each pre-segmented geographical term in the aforementioned pre-segmented geographical term set, in response to the pre-segmented geographical term satisfying the first preset condition, to obtain segmented geographical terms. The first preset condition may be that the pre-segmented geographical term is a compound term or a directional modifier term.

[0031] As an example, the above phrase "Beijing / Chaoyang District / and / north of the Yellow River" can be divided into "Beijing / City / Chaoyang / District / and / north of the Yellow River".

[0032] The third step is to combine the remaining pre-segmented geographical terms and the segmented geographical terms into a set of geographical terms.

[0033] The fourth step involves selecting geographical terms representing logical connections from the aforementioned set of geographical terms, and generating corresponding virtual separators based on the logical connections represented by the selected geographical terms. For example, the virtual separator "∧" can be generated for the phrase "and".

[0034] The fifth step involves replacing the generated logical delimiters in the above geographical terminology set to generate a nested geographical terminology set with a nested logical structure. For example, this nested geographical terminology set could be "Beijing / City / Chaoyang / District ∧ North of the Yellow River".

[0035] The sixth step is to classify each nested geographical term in the above nested geographical term set to generate classified geographical terms.

[0036] The seventh step is to cluster the generated geographical terms after classification to generate a set of text feature information.

[0037] Optionally, after step 102, the following steps are also included:

[0038] The first step is to select textual feature information representing spatial relationship entities from the above set of textual feature information as target feature information.

[0039] In some embodiments, the execution entity may select text feature information representing spatial relationship entities from the set of text feature information as target feature information.

[0040] The second step is to perform contextual error correction on the above target feature information to generate corrected target feature information.

[0041] In some embodiments, the executing entity may perform contextual error correction processing on the target feature information to generate corrected target feature information. This contextual error correction processing may involve completing demonstrative pronouns and omitted words.

[0042] As an example, "the area" can be changed to the area mentioned above, based on the context. As yet another example, "surrounding area" can be replaced with a specific distance radius.

[0043] In practice, the above target feature information can be processed for contextual error correction through the following sub-steps:

[0044] The first sub-step involves matching the categorized geographic terms representing directions, as included in the target feature information, with pre-stored geographic entity topology data. This pre-stored topology data can be a pre-stored topology map containing the categorized geographic terms representing directions. For example, the categorized geographic terms representing directions could be "East," "South," "Upstream," or "Left Bank," etc.

[0045] The second sub-step involves supplementing the categorized geographic terms representing the direction based on the pre-stored geographic entity topological relationship data obtained through matching, resulting in supplemented geographic terms. In practice, topological terms corresponding to the categorized geographic terms representing the direction can be selected from the corresponding pre-stored geographic entity topological relationship data, and then used to replace the categorized geographic terms representing the direction.

[0046] The third sub-step, in response to the existence of fuzzy distance representations among the categorized geographical terms included in the aforementioned target feature information, generates a distance threshold corresponding to each fuzzy distance representation based on a pre-trained regional density model. For example, the fuzzy distance representation could be terms like "nearby" or "a distance away."

[0047] The fourth sub-step involves replacing the corresponding classified geographical terms representing fuzzy distances based on the generated distance threshold.

[0048] The fifth sub-step involves determining the target feature information after the supplementation and replacement processes have been performed as the corrected target feature information.

[0049] In the process of adopting technical solutions to solve the technical problems mentioned in the background, the following technical problems often arise: when generating a similarity retrieval database, only landmarks and locations are stored, without associating terrain, landmarks, buildings and features, resulting in the inability to retrieve the location queried by the user.

[0050] In some optional implementations of certain embodiments, the aforementioned pre-stored geographic entity topology relationship data may be generated through the following steps:

[0051] The first step is to obtain a multimodal geographic data set corresponding to the administrative geographic region. This multimodal geographic data set may include one of the following: orthophotos, elevation data, oblique photogrammetry models, laser point cloud data, topographic maps, and panoramic photographs. It may also include underground pipeline exploration data and geotechnical drilling data. The administrative geographic region may be a pre-defined administrative region.

[0052] The second step involves performing semantic segmentation on at least one orthophoto to create an image mask, resulting in a masked image set. This semantic segmentation can be achieved by using a semantic segmentation model to segment and mask the orthophoto. For example, the semantic segmentation model could be the U-Net model. The masked image set can include building masks and road masks.

[0053] The third step involves performing object detection processing on at least one panoramic photograph included in the aforementioned multimodal geographic data set to generate a landmark information group. In practice, an object detection model can be used to perform object detection processing on at least one panoramic photograph included in the aforementioned multimodal geographic data set to generate the landmark information group. As an example, the object detection model can be a YOLO model or a Faster R-CNN model.

[0054] The fourth step involves clustering the laser point cloud data using a density clustering algorithm to generate clustered point cloud datasets. Based on these clustered datasets, ground feature data is then determined. This ground feature data may include trees and buildings. The density clustering algorithm may be the DBSCAN algorithm.

[0055] The fifth step involves performing ground plane fitting based on the random sample consensus algorithm to obtain ground data. This ground data can include both ground and non-ground points.

[0056] The sixth step involves converting the aforementioned oblique photogrammetry model to generate a 3D building model, and then generating building data from this 3D model. This building data includes the number of building floors and the roof slope. MeshLab can be used to convert the oblique photogrammetry model into a 3D building model and extract the building data from it.

[0057] Step 7: Based on the above-mentioned mask image group, landmark information group, ground feature data, ground data and the above-mentioned building data, generate a joint index of geographic information spatial attributes as pre-stored geographic entity topological relationship data.

[0058] The first to sixth sub-steps described above, as an inventive point of this disclosure, combined with step 106 below, solve the technical problem: "When generating the similarity retrieval database, only landmarks and locations are stored, without associating terrain, landmarks, buildings, and features, resulting in the inability to retrieve the location queried by the user." The reason for the inability to retrieve the location queried by the user is as follows: When generating the similarity retrieval database, only landmarks and locations are stored, without associating terrain, landmarks, buildings, and features, resulting in the inability to retrieve the location queried by the user. If the above factors are resolved, the effect of not being able to retrieve the location queried by the user can be avoided. To achieve this effect, this disclosure firstly obtains a multimodal geographic data information set corresponding to an administrative geographic region. This allows the determination of the multimodal geographic data of the administrative region. Secondly, semantic segmentation processing is performed on at least one orthophoto to perform image masking, obtaining a masked image group. This allows masking of buildings and roads, obtaining masked images. Third, target detection processing is performed on at least one panoramic photograph included in the aforementioned multimodal geographic data information set to generate a landmark information group; based on a density clustering algorithm, laser point cloud data is clustered to generate a clustered point cloud data set, and based on the clustered point cloud data set, ground feature data is determined. This allows for the identification of ground features and landmarks within an administrative region. Fourth, based on a random sampling consistency algorithm, ground plane fitting is performed to obtain ground data; the aforementioned oblique photogrammetry model is transformed to generate a 3D building model, and building data is generated using the 3D building model; based on the aforementioned mask image group, landmark information group, ground feature data, ground data, and building data, a geographic information spatial attribute joint index is generated as pre-stored geographic entity topological relationship data. This allows for the association of terrain, landmarks, buildings, and ground features to generate topological relationship data, enabling the simultaneous retrieval of associated information when a specific data point is found, thus avoiding situations where the user's queried location cannot be found.

[0059] Step 103: Obtain the preset generation rules and generate a spatial operation sequence based on the preset generation rules.

[0060] In some embodiments, the execution entity can obtain preset generation rules and generate a spatial operation sequence based on these preset generation rules. The preset generation rules can be pre-defined rules for sorting spatial operations. For example, the preset generation rules could be: spatial range limitation operation > attribute filtering operation.

[0061] Step 104: Based on the spatial operation sequence, perform retrieval preprocessing on the spatial operation sequence to generate a preprocessed spatial operation sequence.

[0062] In some embodiments, the execution entity may perform retrieval preprocessing on the spatial operation sequence based on the spatial operation sequence to generate a preprocessed spatial operation sequence.

[0063] Step 105: Based on the preprocessed spatial operation sequence, perform stage retrieval processing to generate geographic entity information.

[0064] In some embodiments, the aforementioned execution entity may perform stage retrieval processing based on the aforementioned preprocessed spatial operation sequence to generate geographic entity information.

[0065] In practice, the phased retrieval process can be performed through the following steps:

[0066] The first step is to perform an index retrieval operation based on the spatial semantic index to generate at least one retrieval result.

[0067] In the process of adopting technical solutions to address the technical problems mentioned in the background, the following technical issues often arise: when retrieving geographic entities, a large amount of data usually needs to be retrieved, which takes a considerable amount of time. The conventional solution to this problem is generally to use a traditional inverted index for retrieval. However, this conventional solution still has the following problems: when using a traditional inverted index for retrieval, it is impossible to store and retrieve high-dimensional data; moreover, the traditional inverted index needs to calculate the similarity of the entire dataset, resulting in a long retrieval time; and the retrieved results may not be consistent with the user's needs, causing the user to perform another search, wasting retrieval computing and transmission resources.

[0068] In practice, the above spatial semantic index can be obtained through the following sub-steps:

[0069] The first sub-step is to create a spatial index tree to divide geographic entities into hierarchical structures according to a spatial grid.

[0070] The second sub-step involves constructing a semantic graph index to define the semantic relationship types between geographic entities. These semantic relationship types include: membership, proximity, and functional association.

[0071] The third sub-step involves generating a vector inverted index to vectorize the multimodal features of geographic entities. These multimodal features include spatial coordinates, business attribute labels, and user behavior data.

[0072] The fourth sub-step involves generating an index collaborative triggering mechanism based on preset index configuration information.

[0073] The fifth sub-step involves combining the aforementioned spatial index tree, semantic graph index, vector inverted index, and index collaborative triggering mechanism into a spatial semantic index.

[0074] The sixth sub-step involves performing a secondary filtering of the results returned by the spatial index tree based on the vector inverted index, in response to the number of results returned by the spatial index tree exceeding a preset threshold.

[0075] The first to sixth sub-steps described above, as an inventive point of this disclosure, combined with step 106 below, solve the technical problem: "When retrieving geographic entities, it is usually necessary to retrieve a large amount of data, which takes a long time." The reasons for this long retrieval time are as follows: When retrieving geographic entities, it is usually necessary to retrieve a large amount of data, which takes a long time. Solving these factors can reduce the time required to retrieve geographic entities. To achieve this effect, this disclosure firstly creates a spatial index tree to divide geographic entities into hierarchical structures according to a spatial grid. This allows for the hierarchical representation of each geographic entity. Secondly, it constructs a semantic graph index to define the semantic relationship types between geographic entities. This allows for the determination of semantic relationships through the semantic graph index. Thirdly, it generates a vector inverted index to vectorize and encode the multimodal features of geographic entities. This allows for the encoding of multimodal, multidimensional data. Fourthly, based on preset index configuration information, it generates an index collaborative triggering mechanism; combining the spatial index tree, the semantic graph index, the vector inverted index, and the index collaborative triggering mechanism into a spatial semantic index. Therefore, a spatial semantic index including a vector inverted index can be generated. Fifth, in response to the number of results returned by the spatial index tree exceeding a preset threshold, a secondary filtering is performed on the results returned by the spatial index tree based on the vector inverted index. Thus, the vector inverted index can be executed when triggered. Combined with step "Step 106" below, a geographic entity page corresponding to the geographic entity information is generated, and the geographic entity page is sent to the target user terminal for display. Therefore, the similarity calculation can be performed only on the candidate set using the vector inverted index, avoiding brute-force searching of the entire dataset, thereby avoiding waste of transmission resources and reducing the time users spend retrieving geographic entities.

[0076] The second step involves performing stage filtering on at least one of the above search results to generate a filtered search result sequence, which serves as geographic entity information. This stage filtering can be performed through the following sub-steps:

[0077] The first sub-step involves performing spatial filtering on at least one of the search results to generate a spatially filtered search result set. This spatial filtering can be based on spatial information represented by the target geographic text information. For example, search results within a 3-kilometer radius can be selected.

[0078] The second sub-step involves performing time filtering on the spatially filtered search result set to generate filtered search results. This spatial filtering can be based on the time information represented by the target geographic text information. For example, search results that are open at the current time can be selected.

[0079] The third sub-step involves sorting the filtered search results according to a preset sorting rule to generate a sequence of filtered search results, which serves as geographic entity information. The preset sorting rule can be a pre-defined rule for sorting the search results.

[0080] Step 106: Generate a geographic entity page with corresponding geographic entity information, and send the geographic entity page to the target user terminal for display.

[0081] In some embodiments, the aforementioned executing entity may generate a geographic entity page corresponding to the aforementioned geographic entity information, and send the geographic entity page to the aforementioned target user terminal for display. In practice, the filtered search results in the sequence of filtered search results included in the aforementioned geographic entity information may be displayed in a preset geographic entity page template in the order they appear in the sequence to generate a geographic entity page.

[0082] The above embodiments of this disclosure have the following beneficial effects: the geographic information retrieval method based on a large language model in some embodiments of this disclosure avoids the waste of transmission resources. Specifically, the waste of transmission resources is caused by the inability to process fuzzy statements input by users, resulting in a large number of results that do not match the user's search purpose, which leads to the need to consume a lot of transmission resources when transmitting the results to the user, thus wasting transmission resources. Based on this, the geographic information retrieval method based on a large language model in some embodiments of this disclosure first obtains the target geographic text information input by the target user terminal. Thus, the query information input by the user can be determined. Second, based on a semantic parser, the above target geographic text information is identified and processed to generate a text feature information set. Thus, the text features can be parsed by the semantic parser. Then, a preset generation rule is obtained, and a spatial operation sequence is generated based on the above preset generation rule. Thus, the retrieval order can be determined according to the preset rule. Afterward, based on the above spatial operation sequence, the spatial operation sequence is preprocessed to generate a preprocessed spatial operation sequence. Thus, retrieval preprocessing can be performed before retrieval, thereby concretizing fuzzy statements. Next, based on the preprocessed spatial operation sequence described above, a stage retrieval process is performed to generate geographic entity information. Thus, a retrieval can be performed using specific query information, generating geographic entity information that matches the user's query. Finally, a geographic entity page corresponding to the aforementioned geographic entity information is generated and sent to the target user terminal for display. This ensures accurate search results are sent to the user, avoiding the generation of numerous useless search results and thus preventing the waste of transmission resources.

[0083] Further reference Figure 2 As an implementation of the methods shown in the above figures, this disclosure provides some embodiments of a geographic information retrieval device based on a large language model. These device embodiments are similar to... Figure 1 Corresponding to the method embodiments shown, this geographic information retrieval device based on a large language model can be specifically applied to various electronic devices.

[0084] like Figure 2As shown, some embodiments of the geographic information retrieval device 200 based on a large language model include: a first acquisition unit 201, an identification unit 202, a second acquisition unit 203, a retrieval preprocessing unit 204, a stage retrieval unit 205, and a generation unit 206. The system comprises the following components: a first acquisition unit 201, configured to acquire target geographic text information input by a target user terminal; an identification unit 202, configured to perform identification processing on the target geographic text information based on a semantic parser to generate a text feature information set, wherein the text feature information in the text feature information set is one of the following: spatial relationship entities, geographic attribute information, and logical relationship information; a second acquisition unit 203, configured to acquire preset generation rules and generate a spatial operation sequence based on the preset generation rules; a retrieval preprocessing unit 204, configured to perform retrieval preprocessing on the spatial operation sequence based on the spatial operation sequence to generate a preprocessed spatial operation sequence; a stage retrieval unit 205, configured to perform stage retrieval processing based on the preprocessed spatial operation sequence to generate geographic entity information; and a generation unit 206, configured to generate a geographic entity page corresponding to the geographic entity information and send the geographic entity page to the target user terminal for display.

[0085] It is understandable that the units and references recorded in the geographic information retrieval device 200 based on the large language model are... Figure 1 The steps in the described method correspond to each other. Therefore, the operations, features, and beneficial effects described above for the method are also applicable to the geographic information retrieval device 200 based on a large language model and the units contained therein, and will not be repeated here.

[0086] The following is for reference. Figure 3 This document illustrates a structural schematic of an electronic device 300 suitable for implementing some embodiments of the present disclosure. The electronic devices in some embodiments of the present disclosure may include, but are not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of this disclosure.

[0087] like Figure 3As shown, the electronic device 300 may include a processing unit 301 (e.g., a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 into a random access memory (RAM) 303. The RAM 303 also stores various programs and data required for the operation of the electronic device 300. The processing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0088] Typically, the following devices can be connected to I / O interface 305: input devices 306 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 307 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 308 including, for example, magnetic tapes, hard disks, etc.; and communication devices 309. Communication device 309 allows electronic device 300 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 3 An electronic device 300 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively. Figure 3 Each box shown can represent a device or multiple devices as needed.

[0089] In particular, according to some embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, some embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 309, or installed from storage device 308, or installed from ROM 302. When the computer program is executed by processing device 301, it performs the functions defined in the methods of some embodiments of this disclosure.

[0090] It should be noted that, in some embodiments of this disclosure, the computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In some embodiments of this disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of this disclosure, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0091] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0092] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire target geographic text information input by a target user terminal; perform recognition processing on the aforementioned target geographic text information based on a semantic parser to generate a text feature information set, wherein the text feature information in the aforementioned text feature information set is one of the following: spatial relationship entities, geographic attribute information, and logical relationship information; acquire preset generation rules, and generate a spatial operation sequence based on the aforementioned preset generation rules; perform retrieval preprocessing on the aforementioned spatial operation sequence based on the aforementioned spatial operation sequence to generate a preprocessed spatial operation sequence; perform stage retrieval processing based on the aforementioned preprocessed spatial operation sequence to generate geographic entity information; generate a geographic entity page corresponding to the aforementioned geographic entity information, and send the aforementioned geographic entity page to the aforementioned target user terminal for display.

[0093] Computer program code for performing operations of some embodiments of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0094] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0095] The units described in some embodiments of this disclosure can be implemented in software or hardware. The described units can also be housed in a processor; for example, a processor may be described as including a first acquisition unit, an identification unit, a second acquisition unit, a retrieval preprocessing unit, a stage retrieval unit, and a generation unit. The names of these units do not necessarily limit the specific unit; for example, the first acquisition unit may also be described as "a unit that acquires target geographic text information input by a target user terminal."

[0096] The functions described above in this document can be performed, at least in part, by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: Field Programmable Gate Arrays (FPGAs), Application-Specific Integrated Circuits (ASICs), Application Standard Products (ASSPs), System-on-Chip (SoCs), Complex Programmable Logic Devices (CPLDs), and so on.

[0097] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A geographic information retrieval method based on a large language model, comprising: Obtain the target geographic text information input by the target user terminal; Based on a semantic parser, the target geographic text information is identified and processed to generate a text feature information set, wherein the text feature information in the text feature information set is one of the following: spatial relationship entities, geographic attribute information, and logical relationship information; The step of using a semantic parser to identify and process the target geographic text information to generate a text feature information set includes: The target geographic text information is subjected to geographic term pre-segmentation to generate pre-segmented geographic terms, resulting in a pre-segmented geographic term set. For each pre-segmented geographical term in the pre-segmented geographical term set, in response to the pre-segmented geographical term satisfying a first preset condition, the pre-segmented geographical term is subjected to a secondary segmentation process to obtain segmented geographical terms; Combine the remaining pre-segmented geographical terms and the segmented geographical terms into a set of geographical terms; From the set of geographical terms, select each geographical term that represents a logical connection, and generate the corresponding virtual delimiter based on the logical connection represented by each selected geographical term. Based on the generated virtual delimiters, the set of geographical terms is replaced to generate a set of geographical terms with a nested logical structure, thus obtaining a nested set of geographical terms. For each nested geographical term in the nested geographical term set, the nested geographical terms are classified to generate classified geographical terms; The generated geographical terms are clustered to generate a set of text feature information. Obtain a preset generation rule, and generate a spatial operation sequence based on the preset generation rule, wherein the spatial operation sequence is used to include the arrangement order of each spatial operation, and the spatial operations in the spatial operation sequence include: spatial range limitation and attribute filtering; Based on the spatial operation sequence, the spatial operation sequence is preprocessed to generate a preprocessed spatial operation sequence; Based on the preprocessed spatial operation sequence, a stage retrieval process is performed to generate geographic entity information; Generate a geographic entity page corresponding to the geographic entity information, and send the geographic entity page to the target user terminal for display.

2. The method according to claim 1, wherein, After the semantic parser is used to identify and process the target geographic text information to generate a text feature information set, the method further includes: Text feature information representing spatial relationship entities is selected from the set of text feature information and used as target feature information; The target feature information is subjected to contextual error correction processing to generate corrected target feature information.

3. The method according to claim 2, wherein, The step of performing contextual error correction processing on the target feature information to generate corrected target feature information includes: In response to the presence of categorized geographic terms representing directions in the target feature information, the categorized geographic terms representing directions are matched with pre-stored geographic entity topology relationship data. Based on the pre-stored topological relationship data of geographic entities obtained through matching, the classified geographic terms representing the direction are supplemented to obtain the supplemented geographic terms. In response to the existence of fuzzy distance representations in the categorized geographical terms included in the target feature information, for each categorized geographical term representing fuzzy distance, a distance threshold corresponding to the categorized geographical term is generated based on a pre-trained regional density model. Based on the generated distance threshold, the corresponding classified geographical terms representing fuzzy distance are replaced. The target feature information that has undergone supplementation and replacement processing is identified as the corrected target feature information.

4. The method according to claim 1, wherein, The step of performing a stage retrieval process based on the preprocessed spatial operation sequence to generate geographic entity information includes: Based on spatial semantic indexing, perform index retrieval operations to generate at least one retrieval result; A phase filtering is performed on the at least one search result to generate a filtered search result sequence as geographic entity information.

5. The method according to claim 4, wherein, The step of performing phase filtering on the at least one search result to generate a filtered search result sequence as geographic entity information includes: Spatial filtering is performed on the at least one search result to generate a spatially filtered search result set; The spatially filtered search result set is subjected to time filtering to generate filtered search results; The filtered search results are sorted according to a preset sorting rule to generate a sequence of filtered search results, which serves as geographic entity information.

6. A geographic information retrieval device based on a large language model, comprising: The first acquisition unit is configured to acquire target geographic text information input by the target user terminal; The identification unit is configured to perform identification processing on the target geographic text information based on a semantic parser to generate a text feature information set, wherein the text feature information in the text feature information set is one of the following: spatial relationship entities, geographic attribute information, and logical relationship information; the identification unit is further configured to: The target geographic text information is subjected to geographic term pre-segmentation to generate pre-segmented geographic terms, resulting in a pre-segmented geographic term set. For each pre-segmented geographical term in the pre-segmented geographical term set, in response to the pre-segmented geographical term satisfying a first preset condition, the pre-segmented geographical term is subjected to a secondary segmentation process to obtain segmented geographical terms; Combine the remaining pre-segmented geographical terms and the segmented geographical terms into a set of geographical terms; From the set of geographical terms, select each geographical term that represents a logical connection, and generate the corresponding virtual delimiter based on the logical connection represented by each selected geographical term. Based on the generated virtual delimiters, the set of geographical terms is replaced to generate a set of geographical terms with a nested logical structure, thus obtaining a nested set of geographical terms. For each nested geographical term in the nested geographical term set, the nested geographical terms are classified to generate classified geographical terms; The generated geographical terms are clustered to generate a set of text feature information. The second acquisition unit is configured to acquire a preset generation rule and generate a spatial operation sequence based on the preset generation rule, wherein the spatial operation sequence is used to include the arrangement order of each spatial operation, and the spatial operations in the spatial operation sequence include: spatial range limitation and attribute filtering; The retrieval preprocessing unit is configured to perform retrieval preprocessing on the spatial operation sequence based on the spatial operation sequence to generate a preprocessed spatial operation sequence; The stage retrieval unit is configured to perform stage retrieval processing based on the preprocessed spatial operation sequence to generate geographic entity information; The generation unit is configured to generate a geographic entity page corresponding to the geographic entity information and send the geographic entity page to the target user terminal for display.

7. An electronic device, comprising: One or more processors; A storage device on which one or more programs are stored; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1 to 5.

8. A computer-readable medium having a computer program stored thereon, wherein, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 5.

Images