A library self-service navigation method, device, equipment and medium

Abstract

This application relates to a library self-service navigation method, device, equipment, and medium. The method includes: acquiring multi-source heterogeneous data from the library, and performing data cleaning and structuring processing on the multi-source heterogeneous data to obtain structured multi-source data; extracting spatial features and performing semantic association analysis on the structured multi-source data to obtain environmental feature data; constructing topological relationships and assigning weights to the environmental feature data to construct a navigation topology map; based on user-input navigation request data, performing optimal path search under multiple constraints within the navigation topology map to obtain initial navigation path data; and segmenting the initial navigation path data and generating turning instructions to obtain the path navigation result. This method enables end-to-end coherent navigation from resource semantics to spatial paths, effectively improving the intelligence level and user experience of library self-service navigation.

Description

Technical Field

[0001] This invention belongs to the field of library navigation technology, and in particular relates to a library self-service navigation method, device, equipment and medium. Background Technology

[0002] With the rapid development of smart libraries and self-service technologies, digital navigation systems have become a key means of enhancing the reader experience. Currently, this field mainly relies on the initial integration of electronic maps and library management systems, such as displaying library floor plans via mobile devices and providing keyword-based catalog search services to guide readers to the general area of ​​their target bookshelves.

[0003] In traditional technologies, navigation typically involves a two-step process: first, the user obtains the category number or area name of the target resource through a search system; then, the user must interpret a static site map to plan and find a route to that area. This approach separates resource location from spatial route planning, resulting in discontinuous and fragmented navigation information.

[0004] However, current navigation methods have significant problems: First, information integration is low, with a lack of effective correlation between resource semantics (such as books and equipment) and architectural spatial coordinates, failing to provide end-to-end coherent guidance "from the current location to the specific destination"; second, route planning is rigid, failing to consider real-time environmental factors (such as pedestrian flow and facility status) and personalized user needs, resulting in low guidance efficiency and a poor user experience; third, navigation results are not intuitive, mostly consisting of abstract maps and text, failing to generate step-by-step action instructions that align with human cognition. Therefore, existing technologies are insufficient to meet readers' needs for one-stop, intelligent, accurate, and smooth self-service navigation. Summary of the Invention

[0005] Therefore, it is necessary to provide a library self-service navigation method, device, equipment, and medium to address the aforementioned technical problems.

[0006] Firstly, this application provides a library self-service navigation method, including:

[0007] S1. Acquire multi-source heterogeneous data from the library, and perform data cleaning and structuring on the multi-source heterogeneous data to obtain structured multi-source data; among which, multi-source heterogeneous data includes floor plan data, collection catalog data, and service facility list data.

[0008] S2. Spatial feature extraction and semantic association analysis are performed on structured multi-source data to obtain environmental feature data;

[0009] S3. Construct topological relationships and assign weights to environmental feature data to obtain a navigation topology map;

[0010] S4. Based on the navigation request data input by the user, perform optimal path search under multiple constraints within the navigation topology map to obtain initial navigation path data;

[0011] S5. The initial navigation path data is segmented and described, and turning instructions are generated to obtain the path navigation result; the path navigation result is used to guide users in self-navigation within the library.

[0012] In one embodiment, S1 includes:

[0013] S11. Identify and correct layer misalignment, broken line segments, and coordinate offset errors in the floor plan data to obtain standard vector map data;

[0014] S12. Convert all elements in the standard vector map data to the preset library global spatial coordinate system to obtain unified coordinate map data;

[0015] S13. Based on preset entity recognition rules, extract and classify entities from the unified coordinate map data to obtain spatial entity data; wherein, the spatial entity data includes the location of walls, passages, rooms, service facilities, and the outline information of service facilities;

[0016] S14. Perform non-empty value verification and encoding on the catalog data and service facility list data to obtain standard list data;

[0017] S15. Spatial correlation and attribute fusion are performed on spatial entity data and standard list data to obtain structured multi-source data.

[0018] In one embodiment, S2 includes:

[0019] S21. Identify passable areas and extract key nodes from spatial entity data to obtain node channel network data;

[0020] S22. Calculate the proximity between each node in the node channel network data and the preset set of semantic interest points to obtain node semantic attachment data;

[0021] S23. Calculate the semantic correlation between nodes in the semantic attachment data of nodes, and construct the initial semantic network;

[0022] S24. The semantic correlation between nodes in the initial semantic network is used as an additional attribute of the corresponding edge in the node channel network data to obtain environmental feature data.

[0023] In one embodiment, S3 includes:

[0024] S31. Abstract each key node in the node channel network data into a topology node to obtain a node set;

[0025] S32. For any two nodes in the node set that have a directly connected physical channel in the node channel network data, establish a topological edge to obtain the edge set;

[0026] S33. Perform a multi-dimensional weight calculation on each edge in the edge set to obtain the overall weight of the edge; the expression for the overall weight is:

[0027]

[0028] in, The overall weight of the edges, It is a node With nodes The physical distance between them; It is the maximum physical distance among all edges, used for normalization; It is the semantic relevance of edges obtained from environmental feature data; It is the estimated passage cost of the edge obtained based on historical data or real-time information. The passage cost is used to reflect the flow of people or the status of facilities. It is the maximum value of the estimated passage cost; These are preset weighting coefficients, and ;

[0029] S34. Generate a navigation topology map based on the node set, edge set, and comprehensive weight.

[0030] In one embodiment, S4 includes:

[0031] S41. Perform natural language parsing and keyword matching on the navigation request data input by the user to obtain the target semantic identifier and a set of constraints; the set of constraints includes path preferences and willingness to use service facilities;

[0032] S42. Based on the target semantic identifier, search and locate in the standard list data, determine the target node, and obtain the starting node corresponding to the user's current location;

[0033] S43. Based on the set of constraints, dynamically adjust the comprehensive weights of the relevant edges in the navigation topology graph to obtain the adjusted weights, and update the navigation topology graph based on the adjusted weights to obtain the updated navigation topology graph.

[0034] S44. Starting from the starting node corresponding to the user's current location and ending at the target node, perform a heuristic graph search on the updated navigation topology to obtain the initial navigation path data.

[0035] In one embodiment, S5 includes:

[0036] S51. Identify key turning points in the continuous node sequence in the initial navigation path data to obtain the path segmentation point sequence;

[0037] S52. Based on the path segmentation point sequence, the initial navigation path data is divided into several continuous path segments to obtain a path segment set.

[0038] S53. Combining standard vector map data, calculate the direction of travel and distance for each path segment in the path segment set to obtain the original guidance description for each path segment;

[0039] S54. Perform natural language template filling and conversational conversion on the original guidance description to generate route navigation results.

[0040] In one embodiment, S53 includes:

[0041] S61. Based on the starting and ending coordinates of the path segments in the path segment set, calculate the direction angle to obtain the azimuth data of the path segments.

[0042] S62. Match the path segment orientation data with the preset basic direction to determine the main direction of travel and obtain the direction description data.

[0043] S63. Based on the scale information in the standard vector map data, calculate the actual distance of the path segment to obtain distance measurement data;

[0044] S64. Combining the environmental features of the path segments in the path segment set, identify key landmarks and reference objects within the path segments to obtain landmark identification data;

[0045] S65. Based on direction description data, distance measurement data, and landmark recognition data, the original guidance description for each path segment is obtained.

[0046] Secondly, this application also provides a library self-service navigation device, comprising:

[0047] The data acquisition and preprocessing module is used to acquire multi-source heterogeneous data from the library and perform data cleaning and structuring on the multi-source heterogeneous data to obtain structured multi-source data; among which, multi-source heterogeneous data includes floor plan data, collection catalog data, and service facility list data.

[0048] The feature extraction and semantic analysis module is used to perform spatial feature extraction and semantic association analysis on structured multi-source data to obtain environmental feature data;

[0049] The navigation topology graph construction module is used to construct topological relationships and assign weights to environmental feature data to generate a navigation topology graph.

[0050] The initial path navigation module is used to perform optimal path search under multiple constraints within the navigation topology map based on the navigation request data input by the user, and obtain the initial navigation path data.

[0051] The path navigation result generation module is used to segment and describe the initial navigation path data and generate turning instructions to obtain the path navigation result; the path navigation result is used to guide users in self-navigation within the library.

[0052] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect.

[0053] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0054] The aforementioned library self-service navigation method, device, equipment, and medium achieve deep binding of library resources, service facilities, and architectural spatial coordinates through cleaning, integration, and semantic association of multi-source heterogeneous data. This breaks down the current disconnect between resource positioning and spatial path planning, providing end-to-end coherent navigation guidance. By dynamically adjusting path weights based on user-personalized constraints and real-time environmental information, path planning becomes more flexible and adaptable, improving navigation efficiency. Through key turning point identification, landmark reference integration, and natural language conversion, abstract path information is transformed into intuitive and easy-to-understand step-by-step guidance, aligning with natural human cognitive habits, enhancing the reader's self-service navigation experience, and meeting the library's self-service navigation service needs. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a flowchart illustrating a library self-service navigation method in one embodiment;

[0057] Figure 2 This is a schematic diagram of the structure of a library self-service navigation device in one embodiment. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0059] In one embodiment, reference Figure 1 The document presents a flowchart illustrating the library self-service navigation method provided in this application. This embodiment uses the method applied to a path navigation terminal (hereinafter referred to as the terminal) as an example for explanation. It is understood that this method can also be applied to a server, and further to a system including both a terminal and a server, and is implemented through interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0060] S1. Obtain multi-source heterogeneous data from the library, and perform data cleaning and structuring processing on the multi-source heterogeneous data to obtain structured multi-source data.

[0061] For example, the route navigation terminal obtains multi-source heterogeneous data from different systems in the library through a preset data interface. This multi-source heterogeneous data includes floor plan data, catalog data, and service facility list data.

[0062] Optionally, the floor plan data is derived from vector graphic files in the library's architectural design archives. This includes spatial distribution information of the building structure, such as walls, passageways, doors, windows, and columns, on each floor, as well as preliminary division information for functional areas such as bookshelves and reading areas. The vector graphic files utilize computer-aided design files.

[0063] Optionally, the catalog data comes from the database interface call of the library's book management system, covering the core information of all library resources, including semantic information such as the name, author, classification number, location, and borrowing status of books, periodicals, newspapers, and electronic resources.

[0064] Optionally, the service facility list data is obtained from the real-time synchronization of the library's equipment management system and includes relevant information on all self-service facilities in the library, such as the facility type, installation location, and operating status of self-service borrowing and returning machines, self-service printers, search terminals, water dispensers, and restrooms.

[0065] Optionally, the route navigation terminal performs data cleaning on the acquired multi-source heterogeneous data, including removing redundant data, completing missing data, and correcting erroneous data. For example, it removes duplicate line data from the floor plan, completes missing classification information in the catalogue, and corrects incorrect location coordinates in the service facility list.

[0066] Optionally, the route navigation terminal converts unstructured vector graphics and semi-structured text information into structured data in a unified format. For example, floor plan data is parsed into vector data with spatial coordinates, catalog data is standardized and entered into a preset data model according to fields such as classification number and collection location, and service facility list data is classified according to facility type and associated with spatial coordinates to obtain structured multi-source data with unified format and complete information.

[0067] S2. Spatial feature extraction and semantic association analysis are performed on structured multi-source data to obtain environmental feature data.

[0068] For example, during the spatial feature extraction process, the path navigation terminal uses a convolutional neural network to extract features from the floor plan vector data in the structured multi-source data, identifying and extracting key spatial features such as passage width, turning nodes, stairs, and elevators. Simultaneously, it determines the absolute coordinates and relative positional relationships of each spatial feature, obtaining a set of spatial features. For the catalog data and service facility list data, semantic features are extracted using semantic parsing technology, including the classification semantics of the catalog resources and the functional semantics of the service facilities, to construct a semantic feature library.

[0069] Semantic association analysis is used to establish a mapping relationship between spatial features and semantic features, achieving an effective association between resource semantics and architectural spatial coordinates. Specifically, semantic information such as classification numbers and book titles in the collection catalog data is associated and bound to the spatial coordinates of the corresponding bookshelf areas in the floor plan, clarifying which floor, area, and which bookshelf a particular category of collection resource is located on. Semantic information such as facility type and function in the service facility list data is associated with the spatial coordinates of the corresponding facilities, clarifying the specific spatial location of various self-service facilities, while also associating the semantics of the facility's operational status to ensure synchronous correspondence between semantic and spatial information.

[0070] Optionally, through spatial feature extraction and semantic association analysis, the path navigation terminal integrates environmental feature data containing spatial features, semantic features, and the relationship between the two.

[0071] S3. Construct topological relationships and assign weights to environmental feature data to obtain a navigation topology map.

[0072] For example, the topology is constructed based on the node-edge model in graph theory. The path navigation terminal uses key spatial features and semantically related nodes from the environmental feature data as topology nodes, and traversable paths between nodes as topology edges. The generation of topology edges must be based on the spatial coordinate relationships in the environmental feature data to ensure that the connection of edges conforms to the actual passage logic of the library. Only two nodes that are directly reachable in physical space can generate topology edges, while invalid connections corresponding to inaccessible areas are eliminated. During the weight assignment process, the path navigation terminal determines the weight assignment rules according to the actual usage scenario of the library and the user's navigation needs. Factors affecting the weight include the accessibility of the path, the path length, the relevance of surrounding service facilities, and real-time accessibility. The path navigation terminal integrates the constructed topology nodes, topology edges, and corresponding weight information to generate a navigation topology map. Key spatial feature nodes include stairwells, elevator entrances, passageway intersections, the center point of bookshelf areas, and the location of self-service facilities, etc. Semantically related nodes include nodes of collection resource classification areas and service facility function nodes, etc. Inaccessible areas include the location of walls and columns, etc.

[0073] S4. Based on the navigation request data input by the user, perform optimal path search under multiple constraints within the navigation topology map to obtain initial navigation path data.

[0074] For example, a user can input navigation request data through the interactive interface of the path navigation terminal. The navigation request data includes semantic information of the target resource or type information of the target service facility. The path navigation terminal performs semantic parsing on the navigation request data and maps it to the corresponding target node in the navigation topology map. At the same time, the terminal obtains the user's current spatial location through the built-in positioning module and maps it to the starting node in the navigation topology map.

[0075] Optionally, multiple constraints include user-specific needs constraints and real-time environmental constraints. User-specific needs constraints can be set in advance or selected in real time by the user, such as whether to prefer accessible passages or whether to prioritize the shortest path. Real-time environmental constraints come from the linkage between the path navigation terminal and the library's environmental monitoring system. Information such as real-time pedestrian flow and facility operation status is obtained through image recognition, sensors, etc., such as pedestrian congestion in a certain area, temporary closure of a certain passage, and failure of a certain service facility. This information is then transformed into constraints for path search.

[0076] Optionally, the optimal path search employs a path search algorithm, such as Dijkstra's algorithm. The path navigation terminal, based on the starting and target nodes in the navigation topology graph, and combining multiple constraints and the weight information of each topological edge, traverses and filters paths, eliminating those that do not meet the constraints, and selecting one or more paths with the optimal weights. These are used as the initial navigation path data, which includes the corresponding topological node sequence, topological edge information, and path direction. The interactive interface can be a touchscreen or a voice input module; the semantic information of the target resource includes the book title, author, and classification number; and the type information of the target service facility includes self-service borrowing and returning machines, printers, etc.

[0077] S5. Segment the initial navigation path data and generate turning instructions to obtain the path navigation result.

[0078] For example, during the segmentation description process, the path navigation terminal divides the initial navigation path data into several continuous path segments according to the actual travel scenario based on the distribution of topological nodes in the navigation topology map. Each path segment corresponds to a continuous travel area. The principle of segmentation conforms to the spatial cognition habits of natural people, using obvious spatial feature nodes as segmentation nodes. Spatial feature nodes include stairwells, intersections, bookshelf area markers, etc. Each path segment contains only a single travel direction, which includes going straight, turning, going up or down stairs, etc., avoiding complex path descriptions.

[0079] Optionally, for each path segment, the path navigation terminal generates corresponding segment description information. This description includes the current path's direction of travel and key surrounding spatial features, using a relative position description method to ensure intuitiveness and ease of understanding. Turning instruction generation combines the segment description information and spatial coordinate relationships to generate step-by-step turning instructions that align with natural human movement logic. The generation of turning instructions must match the user's walking rhythm to ensure synchronization with the actual travel scenario. The path navigation terminal integrates the description information of all path segments and the corresponding turning instructions to generate path navigation results. These results can be simultaneously output through both text display and voice broadcast on the terminal. The text display clearly presents the current segment path and the next turning instruction, while the voice broadcast provides real-time reminders based on the user's walking progress and can supplement with semantic prompts for target nodes, guiding users to efficiently complete self-guided navigation within the library and enhancing the user navigation experience.

[0080] In the aforementioned library self-service navigation method, by fusing, cleaning, and associating multi-source heterogeneous data with unified spatial coordinates, the problem of the separation between resource semantics and physical location is effectively solved, achieving end-to-end accurate mapping from user requests to specific spatial targets. By constructing a multi-dimensional weighted navigation topology graph that integrates semantic associations and real-time status, and performing path search under dynamic constraints on this graph, path planning can flexibly adapt to environmental changes and user preferences, improving the intelligence and adaptability of navigation. The path data is transformed into segmented natural language descriptions combining direction, distance, and key landmarks, generating continuous action instructions that conform to human cognitive habits, improving the reader's self-service navigation experience, enhancing the level of smart library self-service, and meeting the reader's self-service navigation needs.

[0081] In an optional embodiment, S1 includes:

[0082] S11. Identify and correct layer misalignment, line breakage, and coordinate offset errors in the floor plan data to obtain standard vector map data.

[0083] Optionally, the floor plan data is sourced from vector graphic files in the library's architectural design archives. Due to storage losses, format conversions, or original drawing deviations, issues such as layer misalignment, broken lines, and coordinate offsets are prone to occur. The path navigation terminal employs a graphic correction algorithm. By extracting feature reference points from the floor plan and comparing them with the actual physical space reference points of the library, a correction mapping relationship is established to correct layer misalignment issues one by one, adjusting misaligned layers to a unified reference. For broken line segments, connectivity detection is performed, and broken line segments are repaired through line endpoint matching and smooth connection processing to ensure the continuity of key structural lines such as passageways and walls. For coordinate offsets, the entire floor plan is calibrated by coordinate translation and rotation using the center point of the library entrance as the reference coordinate to eliminate offset errors and obtain standard vector graphic data.

[0084] S12. Convert all elements in the standard vector map data to the preset library global spatial coordinate system to obtain unified coordinate map data.

[0085] Optionally, the preset global spatial coordinate system of the library is a three-dimensional spatial coordinate system predefined by the path navigation terminal. This coordinate system takes a fixed point on the exterior wall of the library building as its origin and adopts a unified length measurement standard to ensure that all spatial locations within the library can be identified by unique three-dimensional coordinates. The path navigation terminal extracts all graphic elements from the standard vector map data through a coordinate transformation algorithm, converting the original local coordinates of each element into coordinates under the global spatial coordinate system. This achieves coordinate unification of floor plan data for different floors and areas, avoiding subsequent spatial association errors caused by inconsistencies in the coordinate system, and outputting unified coordinate map data with consistent coordinates and complete elements.

[0086] S13. Based on preset entity recognition rules, extract and classify entities from the unified coordinate map data to obtain spatial entity data.

[0087] The spatial entity data includes the location and outline information of walls, passageways, rooms, and service facilities. The preset entity recognition rules are feature matching rules pre-entered by the path navigation terminal. Based on the outline features, size features, and attribute labels of graphic elements, various elements in the unified coordinate map data are identified and classified. For example, walls and rooms are identified by the closure and size range of the outline lines; passageways are identified by the continuity and width features of the lines; and service facilities are identified by matching preset facility outline templates. During the extraction process, the three-dimensional coordinate information and outline parameters of various spatial entities are recorded simultaneously. The extracted entities are then classified and archived to generate spatial entity data.

[0088] S14. Perform non-empty value verification and encoding on the catalog data and service facility list data to obtain standard list data.

[0089] Optionally, the catalog data comes from the database interface of the library's book management system, and the service facility list data comes from real-time synchronization of the library's equipment management system. Both types of data may have issues such as missing fields or inconsistent formats. During the non-empty value verification process, the path navigation terminal checks the core fields of both types of data one by one, removing invalid blank fields, marking missing key fields, and supplementing them with basic default information. The encoding process uses unified encoding rules, such as classification codes and facility type codes, to standardize the text information in both types of data. The catalog is categorized by classification number, and the service facilities are divided by type, resulting in standard list data.

[0090] S15. Spatial correlation and attribute fusion are performed on spatial entity data and standard list data to obtain structured multi-source data.

[0091] Optionally, during the spatial association process, based on the global spatial coordinate system, the path navigation terminal binds the collection catalog and service facilities in the standard list data with the corresponding spatial locations in the spatial entity data. For example, the classification number of a certain type of collection catalog is associated with the three-dimensional coordinates of the corresponding bookshelf area in the spatial entity data, and the information of a certain service facility is associated with the location and outline information of the facility in the spatial entity data. Attribute fusion integrates the attribute information of the two types of data, and merges and archives the coordinate and outline attributes of the spatial entity with the semantic attributes of the standard list to generate structured multi-source data.

[0092] In an optional embodiment, S2 includes:

[0093] S21. Identify passable areas and extract key nodes from spatial entity data to obtain node channel network data.

[0094] Optionally, the spatial entity data includes the location and outline information of walls, passageways, rooms, and service facilities. Accessible area identification is used to distinguish between accessible and inaccessible areas in the entity data. The path navigation terminal employs a region growing algorithm, based on the outline features and attribute information of spatial entities, to remove inaccessible entities such as walls, columns, and non-accessible parts of bookshelves, while retaining accessible areas such as passageways, corridors, elevator lobbies, and stairwells. Simultaneously, connectivity analysis is performed on the accessible areas to ensure that the identified accessible areas conform to the actual passageway logic of the library. Key node extraction is based on the accessible areas. The path navigation terminal selects nodes that are significant for path navigation, including passageway intersections, stairwells, elevator entrances, service facility center points, accessible area entrances and exits, and bookshelf area boundary nodes. The 3D coordinate information and surrounding entity features of each key node are extracted. Key nodes are treated as network nodes, and accessible road segments connecting nodes are treated as network edges, integrating to generate node-passage network data.

[0095] S22. Calculate the proximity between each node in the node channel network data and the preset set of semantic interest points to obtain node semantic attachment data.

[0096] Optionally, a pre-set set of semantic interest points (PIs) is pre-entered and dynamically updated by the path navigation terminal. This set covers all semantically meaningful core targets within the library, including collection resource classification areas, various service facilities, and functional areas. Each PI is associated with corresponding semantic attributes and spatial coordinates. Proximity calculation employs either the Euclidean distance algorithm or the Manhattan distance algorithm. Based on the three-dimensional coordinates of each node in the node channel network data and the three-dimensional coordinates of the PIs, the spatial proximity between them is calculated. Simultaneously, the proximity value is adjusted by incorporating the importance weight of the PIs. Each node is then bound to one or more PIs with the closest proximity and highest correlation. The PI information, proximity value, and association weight of each node are recorded, generating node semantic attachment data and achieving a preliminary association between spatial nodes and semantic information.

[0097] S23. Calculate the semantic correlation between nodes in the semantic attachment data of the nodes, and construct the initial semantic network.

[0098] Optionally, semantic relevance is used to characterize the degree of association between semantic interest points attached to two nodes. The calculation process can employ a cosine similarity semantic similarity algorithm, combined with the classification and functional attributes of the semantic interest points for analysis. For example, nodes corresponding to the same category of collection area have higher semantic relevance than nodes corresponding to different category areas, and nodes corresponding to adjacent service facilities have higher semantic relevance than nodes that are far apart and functionally unrelated. The path navigation terminal traverses all nodes in the node semantic attachment data, calculates the semantic relevance between any two nodes one by one, removes node pairs with extremely low semantic relevance, treats the nodes as semantic nodes, and uses the semantic relevance between nodes as the association strength between nodes, constructing an initial semantic network containing semantic nodes, association strength, and association relationships. The initial semantic network is used to reflect the association logic between the semantic information corresponding to each spatial node.

[0099] S24. The semantic correlation between nodes in the initial semantic network is used as an additional attribute of the corresponding edge in the node channel network data to obtain environmental feature data.

[0100] Optionally, the node channel network data includes the connectivity relationships between nodes and edges at the spatial level, while the initial semantic network includes the semantic association strength between nodes. The fusion of the two can achieve deep binding of spatial and semantic information. Specifically, the path navigation terminal matches the semantic correlation between any two nodes in the initial semantic network to the network edge between the corresponding two nodes in the node channel network data, as an additional semantic attribute of the network edge. At the same time, it retains the original spatial coordinates, passage attributes, and other information in the node channel network, as well as the semantic attachment information corresponding to the nodes, and integrates them to generate environmental feature data containing spatial connectivity relationships, semantic association strength, node attributes, and edge attributes.

[0101] In an optional embodiment, S3 includes:

[0102] S31. Abstract each key node in the node channel network data into a topology node to obtain a node set.

[0103] Optionally, the path navigation terminal abstracts each key node in the node channel network data into a topology node, resulting in a node set. The node channel network data includes key nodes covering spatially significant nodes with navigational characteristics, such as channel intersections, stairwells, elevator entrances, and service facility center points. During the abstraction process, the path navigation terminal retains the core attribute information of each key node, including its three-dimensional spatial coordinates, corresponding semantic attachment information, and node type identifier. It removes details irrelevant to topology construction, such as node contour parameters and redundant attributes, mapping each key node to an independent topology node. All topology nodes are integrated to generate a node set, and each topology node is assigned a unique identifier.

[0104] S32. For any two nodes in the node set that have a directly connected physical channel in the node channel network data, establish a topological edge to obtain the edge set.

[0105] Optionally, the determination of directly connected physical channels is based on the distribution of passable areas in the node channel network data. That is, between the key nodes corresponding to two topological nodes, there exists a continuous, unobstructed, and directly passable channel, such as a corridor between adjacent intersections, or a connecting section between an elevator entrance and a neighboring channel intersection, and the channel does not require transit through other topological nodes. If this condition is met, a topological edge is established between the two topological nodes. Each topological edge is associated with two corresponding topological node identifiers and records the basic information of the physical channel corresponding to the edge. All constructed topological edges are integrated to generate an edge set. The edge set and the node set together constitute the basic framework of the navigation topology map, ensuring that the topological relationships conform to the actual passage logic of the library. The basic channel information includes the channel type and approximate direction.

[0106] S33. Perform a multi-dimensional weight calculation on each edge in the edge set to obtain the overall weight of the edge; the expression for the overall weight is:

[0107]

[0108] in, The overall weight of the edges, It is a node With nodes The physical distance between them; It is the maximum physical distance among all edges, used for normalization; It is the semantic relevance of edges obtained from environmental feature data; It is the estimated passage cost of the edge obtained based on historical data or real-time information. The passage cost is used to reflect the flow of people or the status of facilities. It is the maximum value of the estimated passage cost; These are preset weighting coefficients, and .

[0109] Specifically, the path navigation terminal performs a multi-dimensional weight calculation on each edge in the edge set to obtain the edge's comprehensive weight; in the expression for the comprehensive weight, The weight of the edge represents the priority of the path corresponding to that edge in navigation; the smaller the weight value, the better the path. It is a node With nodes The physical distance between them is derived from the three-dimensional coordinates of two key nodes in the node channel network data and is calculated using a spatial distance algorithm. It is the maximum physical distance among all edges, used to normalize the physical distance, eliminate the influence of different length paths, and ensure that the calculation results of the physical distance dimension are within the same numerical range. It is the semantic relevance of the edges obtained from environmental feature data. It comes from the semantic relevance attribute bound to the node channel network edge and is used to characterize the degree of semantic association between the path and the user's navigation request. It is the estimated passage cost of the edge obtained based on historical data or real-time information. The passage cost is used to reflect the flow of people or the status of facilities. The higher the flow of people or the more inconvenient the passage is due to facility failure / maintenance, the higher the passage cost. Its data comes from the real-time flow data detected by the path navigation terminal and the library environment monitoring system and the statistics of historical navigation passage records. It is the maximum value of the estimated passage cost, used to normalize the passage cost; These are preset weighting coefficients, and The weighting coefficients can be dynamically adjusted according to the library's navigation needs. The path navigation terminal calculates the comprehensive weight of each edge in the edge set according to the above formula, and binds the weight value to the corresponding topological edge.

[0110] S34. Generate a navigation topology map based on the node set, edge set, and comprehensive weight.

[0111] Optionally, the path navigation terminal integrates all topological nodes in the node set, all topological edges in the edge set, and the comprehensive weight corresponding to each topological edge to construct a complete topological structure model. Here, topological nodes correspond to key navigation points within the library, topological edges correspond to traversable paths between points, and the comprehensive weight serves as the core basis for path selection. The navigation topology map also associates the semantic attachment information of each topological node with the basic attributes of each topological edge.

[0112] In an optional embodiment, S4 includes:

[0113] S41. Perform natural language parsing and keyword matching on the navigation request data input by the user to obtain the target semantic identifier and constraint set.

[0114] Optionally, the set of constraints includes path preferences and willingness to use service facilities. Users input navigation request data through the interactive interface of the path navigation terminal. This data is in natural language format. The terminal uses natural language processing technology to parse the data, splitting the request into segments using a word segmentation algorithm, extracting core keywords, and semantically mapping these keywords to obtain a target semantic identifier. This identifier uniquely corresponds to a specific resource or service facility within the library. Simultaneously, keywords representing user needs are extracted and organized to generate a set of constraints. Path preferences include shortest path, accessible path, and path with the least foot traffic. Willingness to use service facilities includes whether the user needs to pass by self-service borrowing / returning machines, search terminals, or other similar facilities.

[0115] S42. Based on the target semantic identifier, search and locate in the standard list data, determine the target node, and obtain the starting node corresponding to the user's current location.

[0116] Optionally, the path navigation terminal, based on the target semantic identifier, searches and locates within the standard list data to determine the target node and obtains the starting node corresponding to the user's current location. The standard list data includes standardized coded catalog data and service facility list data, with each data entry associated with corresponding spatial coordinate information. The path navigation terminal employs a keyword retrieval algorithm to match the target semantic identifier with the coded information and semantic attributes in the standard list data, retrieving the corresponding target resource or service facility. Combining the correlation between spatial entity data and the standard list data, it determines the spatial location of the target resource or service facility and maps it to the corresponding topology node in the navigation topology map, using this topology node as the target node. Simultaneously, the path navigation terminal obtains the user's current three-dimensional spatial coordinates through a built-in positioning module, matching these coordinates to the corresponding topology node in the navigation topology map as the starting node, ensuring the accuracy of the positioning between the starting node and the target node. The positioning module includes Bluetooth positioning, Wi-Fi positioning, etc.

[0117] S43. Based on the set of constraints, dynamically adjust the comprehensive weights of the relevant edges in the navigation topology graph to obtain the adjusted weights, and update the navigation topology graph based on the adjusted weights to obtain the updated navigation topology graph.

[0118] Optionally, the dynamic adjustment process of the original edge weights in the navigation topology graph needs to incorporate adjustment rules based on path preferences and service facility usage intentions within the constraint set. For example, if a user prefers accessible paths, the weight of topology edges passing through accessible passages is appropriately reduced, while the weight of topology edges not passing through accessible passages but containing steps is appropriately increased. If a user intends to use self-service borrowing and returning machines, the weight of topology edges passing through nodes corresponding to these machines is appropriately reduced. During the adjustment process, the fundamental dimension of weight calculation remains unchanged; only the weight values ​​of each topology edge are dynamically corrected according to the constraints. The adjusted weights of each edge are then used to replace the original weights in the navigation topology graph, completing the update and ensuring the updated topology graph aligns with the user's personalized needs.

[0119] S44. Starting from the starting node corresponding to the user's current location and ending at the target node, perform a heuristic graph search on the updated navigation topology to obtain the initial navigation path data.

[0120] Optionally, the heuristic graph search employs the A* algorithm, which combines the convenience of forward search with the accuracy of reverse heuristics, enabling rapid selection of the optimal path. The path navigation terminal uses the starting node as the search origin and the target node as the search endpoint. Based on the adjusted weights of each topological edge in the updated navigation topology graph, a heuristic function is set. By traversing the topological nodes and edges in the navigation topology graph, the path with the optimal weight and meeting the constraints is selected. This path contains core information such as a continuous sequence of topological nodes, topological edge information, and path direction, and is used as the initial navigation path data. The initial navigation path data accurately matches the user's personalized needs while considering path convenience and semantic relevance. The heuristic function is used to estimate the optimal path cost from the current node to the target node.

[0121] In an optional embodiment, S5 includes:

[0122] S51. Identify key turning points in the continuous node sequence in the initial navigation path data to obtain the path segmentation point sequence.

[0123] Optionally, the initial navigation path data includes a continuous sequence of topological nodes, topological edge information, and path direction. The continuous node sequence corresponds to all topological nodes traversed by the user along the complete path from the starting node to the target node. Key turning point identification is used to filter out nodes with significant changes in path direction and navigational characteristics. The path navigation terminal employs a node direction analysis algorithm to identify key turning points such as turning nodes, stair-climbing nodes, and passageway switching nodes by calculating the change in the angle between two adjacent topological edges and combining this with node type identifiers, thus eliminating continuous ordinary nodes in straight sections of the path. During the identification process, the three-dimensional coordinates, node type, and corresponding topological edge information of each key turning point are recorded simultaneously. All key turning points are arranged according to the path travel order to generate a path segmentation point sequence. This sequence provides a clear basis for subsequent path segmentation, ensuring that the segmentation conforms to the actual travel logic.

[0124] S52. Based on the path segmentation point sequence, the initial navigation path data is divided into several continuous path segments to obtain a path segment set.

[0125] Optionally, during the path segmentation process, the path navigation terminal uses two adjacent key turning points in the path segmentation point sequence as the start and end points of a path segment. It then splits the corresponding node sequence, topological edge information, and path direction in the initial navigation path data, generating several continuous and independent path segments. Each path segment corresponds to a single travel scenario, containing only one direction of travel with no change in direction. The segmented path segments are arranged sequentially according to the travel order, forming a path segment set. Simultaneously, the path navigation terminal assigns a unique identifier to each path segment, associates it with the corresponding start and end point turning point information, and records the topological edge set corresponding to the path segment, ensuring that the information of each path segment is complete and traceable.

[0126] S53. Combining standard vector map data, calculate the direction of travel and distance for each path segment in the path segment set to obtain the original guidance description for each path segment.

[0127] Optionally, the standard vector map data includes spatial information such as the building structure and passageway distribution of each floor of the library, along with corresponding coordinate data. The direction of travel is calculated based on the three-dimensional coordinates of the starting and ending points of a path segment. Through coordinate difference analysis, combined with the orientation reference in the standard vector map, the specific direction of travel for that path segment is determined. Distance calculation uses a spatial distance algorithm to calculate the physical distance between the starting and ending points of a path segment. The path navigation terminal integrates the direction of travel, relative distance, and surrounding key spatial features of each path segment to generate an original guidance description, which includes the passage information for the path segment.

[0128] S54. Perform natural language template filling and conversational conversion on the original guidance description to generate route navigation results.

[0129] Optionally, the route navigation terminal has built-in multiple natural language guidance templates tailored to human cognitive habits. These templates cover different travel scenarios such as going straight, turning, and going up and down stairs. Each template has reserved spaces for information such as travel direction, relative distance, and surrounding features. During the filling process, the route navigation terminal fills the core information from the original guidance descriptions of each route segment into the matching natural language templates, replacing the reserved spaces. The conversational conversion does not use spoken language; instead, it transforms the technical language in the original guidance descriptions into easily understandable and concise natural language, eliminating jargon to ensure the guidance content is intuitive and aligns with users' everyday understanding. After conversion, the route navigation terminal integrates the guidance descriptions of all route segments in chronological order, adds semantic prompts for target nodes, and generates a complete route navigation result. The navigation result can be simultaneously output through text display and voice broadcast on the route navigation terminal. The text display clearly presents the guidance for the current route segment and the next travel instructions, while the voice broadcast provides real-time reminders based on the user's walking progress, achieving end-to-end coherent guidance and enhancing the user's self-navigation experience.

[0130] In an optional embodiment, S53 includes:

[0131] S61. Based on the starting and ending coordinates of the path segments in the path segment set, calculate the direction angle to obtain the azimuth data of the path segments.

[0132] Optionally, the path segment set includes the three-dimensional spatial coordinates of the starting point and the turning point of each path segment. The azimuth angle calculation is used to accurately represent the actual travel direction of the path segment. The azimuth angle calculation method is adopted. Based on the horizontal plane coordinates of the starting point and the ending point, and with a preset azimuth benchmark as a reference, the azimuth angle between the two points is calculated clockwise, that is, the angle between the path segment and the azimuth benchmark. The calculated azimuth angle value is used as the azimuth data of the path segment, and the corresponding path segment identifier is synchronously associated to ensure that the azimuth data corresponds one-to-one with the path segment.

[0133] S62. Match the path segment orientation data with the preset basic direction at an angle to determine the main direction of travel and obtain the direction description data.

[0134] Optionally, the route navigation terminal has eight preset standard basic directions: north, northeast, east, southeast, south, southwest, west, and northwest. Each basic direction corresponds to a preset angle range. During the angle matching process, the route navigation terminal compares the direction angles in the route segment's direction data with the angle ranges corresponding to each basic direction one by one, and selects the basic direction with the highest matching degree as the main travel direction for that route segment. If the direction angle is at the intersection of the angle ranges of two basic directions, the terminal combines the standard vector map data associated with the route segment to determine the main travel direction that better fits the actual passage scenario of the library, and converts the determined main travel direction into a standardized direction description to generate direction description data.

[0135] S63. Based on the scale information in the standard vector map data, calculate the actual distance of the path segment to obtain distance measurement data.

[0136] Optionally, the attribute information of the standard vector map data includes preset scale information. This scale is used to represent the correspondence between the distances on the vector map and the actual physical distances within the library, and is read and stored in advance by the path navigation terminal. During the calculation process, the path navigation terminal extracts the distances on the standard vector map between the starting and ending points of the path segment, and converts them in conjunction with the scale information to obtain the actual physical distance of the path segment. An intuitive relative distance description is used, and the converted relative distance description is used as the distance measurement data, associated with the corresponding path segment identifier, to ensure that the distance description is consistent with the actual travel scenario and is easy to understand.

[0137] S64. Combining the environmental features of the path segments in the path segment set, identify key landmarks and reference objects within the path segments to obtain landmark identification data.

[0138] Optionally, each path segment in the path segment set is associated with corresponding environmental feature information, which comes from the topological edge information associated with the path segment and the spatial entity information in the standard vector map data. Key landmark and reference point identification employs a feature matching algorithm. The path navigation terminal pre-loads feature templates of common key landmarks within the library, compares the environmental features of the path segment with the landmark feature templates, identifies key landmarks and reference points within the path segment that are navigational and easily identifiable by users, and simultaneously records the relative positions of the landmarks / reference points with respect to the path segment, generating landmark identification data.

[0139] S65. Based on direction description data, distance measurement data, and landmark recognition data, the original guidance description for each path segment is obtained.

[0140] Optionally, the route navigation terminal integrates the direction description data, distance measurement data, and landmark identification data corresponding to the same route segment, and constructs the content of the original guidance description according to the logical order of preset description rules. The original guidance description is presented in concise and standardized technical language, without colloquial expressions, clearly stating the passage information of the route segment, and associating it with the corresponding route segment identifier, ensuring that the original guidance description information of each route segment is complete and logically clear.

[0141] The aforementioned library self-service navigation method achieves precise correlation between resource semantic information and building spatial coordinates by deeply integrating and structuring multi-source heterogeneous data, thus solving the problem of information fragmentation in traditional methods. It provides end-to-end coherent navigation guidance from the user's current location to the specific target resource. By constructing a navigation topology map that integrates semantic and spatial characteristics and introducing a dynamic weighting mechanism, path planning can flexibly respond to real-time environmental changes and user personalized preferences, overcoming the rigidity of traditional static path planning and improving navigation accuracy and efficiency. Furthermore, by converting optimal path data into natural language instructions containing step-by-step directions, precise distances, and descriptions of key landmarks, it generates navigation results highly aligned with human cognitive habits, enhancing the intuitiveness of guidance and the user's autonomous wayfinding experience, thus achieving a one-stop self-service navigation service.

[0142] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0143] Based on the same inventive concept, this application also provides a library self-service navigation device for implementing the library self-service navigation method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations of one or more embodiments of the library self-service navigation device provided below can be found in the limitations of the library self-service navigation method described above, and will not be repeated here.

[0144] In one exemplary embodiment, such as Figure 2 As shown, a schematic diagram of the structure of a library self-service navigation device 10 is provided, including:

[0145] The data acquisition and preprocessing module 11 is used to acquire multi-source heterogeneous data from the library and perform data cleaning and structuring on the multi-source heterogeneous data to obtain structured multi-source data; among which, the multi-source heterogeneous data includes floor plan data, collection catalog data, and service facility list data.

[0146] The feature extraction and semantic analysis module 12 is used to perform spatial feature extraction and semantic association analysis on structured multi-source data to obtain environmental feature data;

[0147] The navigation topology construction module 13 is used to construct topological relationships and assign weights to environmental feature data to obtain a navigation topology map.

[0148] The initial path navigation module 14 is used to perform optimal path search under multiple constraints within the navigation topology map based on the navigation request data input by the user, and obtain initial navigation path data.

[0149] The path navigation result generation module 15 is used to segment and describe the initial navigation path data and generate turning instructions to obtain the path navigation result; the path navigation result is used to guide users in self-navigation within the library.

[0150] Furthermore, the data acquisition and preprocessing module 11 is also used for:

[0151] S11. Identify and correct layer misalignment, broken line segments, and coordinate offset errors in the floor plan data to obtain standard vector map data;

[0152] S12. Convert all elements in the standard vector map data to the preset library global spatial coordinate system to obtain unified coordinate map data;

[0153] S13. Based on preset entity recognition rules, extract and classify entities from the unified coordinate map data to obtain spatial entity data; wherein, the spatial entity data includes the location of walls, passages, rooms, service facilities, and the outline information of service facilities;

[0154] S14. Perform non-empty value verification and encoding on the catalog data and service facility list data to obtain standard list data;

[0155] S15. Spatial correlation and attribute fusion are performed on spatial entity data and standard list data to obtain structured multi-source data.

[0156] Furthermore, the feature extraction and semantic analysis module 12 is also used for:

[0157] S21. Identify passable areas and extract key nodes from spatial entity data to obtain node channel network data;

[0158] S22. Calculate the proximity between each node in the node channel network data and the preset set of semantic interest points to obtain node semantic attachment data;

[0159] S23. Calculate the semantic correlation between nodes in the semantic attachment data of nodes, and construct the initial semantic network;

[0160] S24. The semantic correlation between nodes in the initial semantic network is used as an additional attribute of the corresponding edge in the node channel network data to obtain environmental feature data.

[0161] Furthermore, the navigation topology map construction module 13 is also used for:

[0162] S31. Abstract each key node in the node channel network data into a topology node to obtain a node set;

[0163] S32. For any two nodes in the node set that have a directly connected physical channel in the node channel network data, establish a topological edge to obtain the edge set;

[0164] S33. Perform a multi-dimensional weight calculation on each edge in the edge set to obtain the overall weight of the edge; the expression for the overall weight is:

[0165]

[0166] in, The overall weight of the edges, It is a node With nodes The physical distance between them; It is the maximum physical distance among all edges, used for normalization; It is the semantic relevance of edges obtained from environmental feature data; It is the estimated passage cost of the edge obtained based on historical data or real-time information. The passage cost is used to reflect the flow of people or the status of facilities. It is the maximum value of the estimated passage cost; These are preset weighting coefficients, and ;

[0167] S34. Generate a navigation topology map based on the node set, edge set, and comprehensive weight.

[0168] Furthermore, the initial path navigation module 14 is also used for:

[0169] S41. Perform natural language parsing and keyword matching on the navigation request data input by the user to obtain the target semantic identifier and a set of constraints; the set of constraints includes path preferences and willingness to use service facilities;

[0170] S42. Based on the target semantic identifier, search and locate in the standard list data, determine the target node, and obtain the starting node corresponding to the user's current location;

[0171] S43. Based on the set of constraints, dynamically adjust the comprehensive weights of the relevant edges in the navigation topology graph to obtain the adjusted weights, and update the navigation topology graph based on the adjusted weights to obtain the updated navigation topology graph.

[0172] S44. Starting from the starting node corresponding to the user's current location and ending at the target node, perform a heuristic graph search on the updated navigation topology to obtain the initial navigation path data.

[0173] Furthermore, the route navigation result generation module 15 is also used for:

[0174] S51. Identify key turning points in the continuous node sequence in the initial navigation path data to obtain the path segmentation point sequence;

[0175] S52. Based on the path segmentation point sequence, the initial navigation path data is divided into several continuous path segments to obtain a path segment set.

[0176] S53. Combining standard vector map data, calculate the direction of travel and distance for each path segment in the path segment set to obtain the original guidance description for each path segment;

[0177] S54. Perform natural language template filling and conversational conversion on the original guidance description to generate route navigation results.

[0178] Furthermore, the route navigation result generation module 15 is also used for:

[0179] S61. Based on the starting and ending coordinates of the path segments in the path segment set, calculate the direction angle to obtain the azimuth data of the path segments.

[0180] S62. Match the path segment orientation data with the preset basic direction to determine the main direction of travel and obtain the direction description data.

[0181] S63. Based on the scale information in the standard vector map data, calculate the actual distance of the path segment to obtain distance measurement data;

[0182] S64. Combining the environmental features of the path segments in the path segment set, identify key landmarks and reference objects within the path segments to obtain landmark identification data;

[0183] S65. Based on direction description data, distance measurement data, and landmark recognition data, the original guidance description for each path segment is obtained.

[0184] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of a library self-service navigation method as described above.

[0185] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0186] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The components described as separate parts may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this disclosure according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0187] The above-described embodiments are merely illustrative of several implementation methods of the embodiments of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of the patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the embodiments of this application, and these modifications and improvements all fall within the protection scope of the embodiments of this application.

Claims

1. A library self-service navigation method, characterized in that, The method includes: S1. Acquire multi-source heterogeneous data from the library, and perform data cleaning and structuring processing on the multi-source heterogeneous data to obtain structured multi-source data; wherein, the multi-source heterogeneous data includes floor plan data, collection catalog data, and service facility list data. S2. Spatial feature extraction and semantic association analysis are performed on the structured multi-source data to obtain environmental feature data; S3. Perform topological relationship construction and weight assignment on the environmental feature data to construct a navigation topology map; S4. Based on the navigation request data input by the user, perform optimal path search under multiple constraints within the navigation topology map to obtain initial navigation path data; S5. The initial navigation path data is segmented and described, and turning instructions are generated to obtain the path navigation result; wherein, the path navigation result is used to guide the user's self-navigation within the library.

2. The method according to claim 1, characterized in that, S1 includes: S11. Identify and correct layer misalignment, line breakage, and coordinate offset errors in the floor plan data to obtain standard vector map data; S12. Convert all elements in the standard vector map data to the preset library global spatial coordinate system to obtain unified coordinate map data; S13. Based on preset entity recognition rules, extract and classify entities from the unified coordinate map data to obtain spatial entity data; wherein, the spatial entity data includes the location of walls, passages, rooms, service facilities, and the outline information of service facilities; S14. Perform non-empty value verification and encoding on the aforementioned collection catalog data and service facility list data to obtain standard list data; S15. Spatial association and attribute fusion are performed on the spatial entity data and the standard list data to obtain the structured multi-source data.

3. The method according to claim 2, characterized in that, S2 includes: S21. The spatial entity data is used to identify passable areas and extract key nodes to obtain node channel network data. S22. Calculate the proximity between each node in the node channel network data and the preset set of semantic interest points to obtain node semantic attachment data; S23. Calculate the semantic correlation between nodes in the node semantic attachment data, and construct the initial semantic network; S24. The semantic relevance between nodes in the initial semantic network is used as an additional attribute of the corresponding edge in the node channel network data to obtain the environmental feature data.

4. The method according to claim 3, characterized in that, S3 includes: S31. Abstract each key node in the node channel network data into a topology node to obtain a node set; S32. For any two nodes in the node set, if there is a physical channel that directly connects them in the node channel network data, establish a topological edge to obtain an edge set. S33. Perform a multi-dimensional weight calculation on each edge in the edge set to obtain the comprehensive weight of the edge; the expression for the comprehensive weight is: in, The overall weight of the edges, It is a node With nodes The physical distance between them; It is the maximum physical distance among all edges, used for normalization; It is the semantic relevance of the edges obtained from the environmental feature data; It is the estimated passage cost of the edge obtained based on historical data or real-time information, and the passage cost is used to reflect the flow of people or the status of facilities; It is the maximum value of the estimated passage cost; These are preset weighting coefficients, and ; S34. Generate the navigation topology map based on the node set, the edge set, and the comprehensive weight.

5. The method according to claim 4, characterized in that, S4 includes: S41. Perform natural language parsing and keyword matching on the navigation request data input by the user to obtain a target semantic identifier and a set of constraints; the set of constraints includes path preferences and willingness to use service facilities; S42. Based on the target semantic identifier, search and locate in the standard list data, determine the target node, and obtain the starting node corresponding to the user's current location; S43. Based on the set of constraints, dynamically adjust the comprehensive weights of the relevant edges in the navigation topology graph to obtain the adjusted weights, and update the navigation topology graph based on the adjusted weights to obtain the updated navigation topology graph. S44. Starting from the starting node corresponding to the user's current location and ending at the target node, perform a heuristic graph search on the updated navigation topology to obtain the initial navigation path data.

6. The method according to claim 5, characterized in that, S5 includes: S51. Identify key turning points in the continuous node sequence in the initial navigation path data to obtain a path segmentation point sequence; S52. Based on the path segmentation point sequence, the initial navigation path data is divided into several continuous path segments to obtain a path segment set; S53. Combining the standard vector map data, calculate the travel direction and distance for each path segment in the path segment set to obtain the original guidance description of each path segment; S54. Perform natural language template filling and colloquialization on the original guidance description to generate the path navigation result.

7. The method according to claim 6, characterized in that, S53 includes: S61. Based on the starting and ending coordinates of the path segments in the set of path segments, calculate the direction angle to obtain the azimuth data of the path segments. S62. Match the path segment orientation data with the preset basic direction by angle to determine the main direction of travel and obtain the direction description data. S63. Based on the scale information in the standard vector map data, calculate the actual distance of the path segment to obtain distance measurement data; S64. Combining the environmental features of the path segments in the path segment set, identify key landmarks and reference objects within the path segments to obtain landmark identification data; S65. Based on the direction description data, the distance measurement data, and the landmark identification data, the original guidance description for each of the path segments is obtained.

8. A library self-service navigation device, characterized in that, The device includes: The data acquisition and preprocessing module is used to acquire multi-source heterogeneous data from the library and perform data cleaning and structuring processing on the multi-source heterogeneous data to obtain structured multi-source data; wherein, the multi-source heterogeneous data includes floor plan data, collection catalog data, and service facility list data; The feature extraction and semantic analysis module is used to perform spatial feature extraction and semantic association analysis on the structured multi-source data to obtain environmental feature data; The navigation topology graph construction module is used to construct topological relationships and assign weights to the environmental feature data to construct a navigation topology graph. The initial path navigation module is used to perform optimal path search under multiple constraints within the navigation topology map based on navigation request data input by the user, and obtain initial navigation path data. The path navigation result generation module is used to segment and describe the initial navigation path data and generate turning instructions to obtain path navigation results; wherein, the path navigation results are used to guide the user's self-navigation within the library.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images