An automated geographic information system mapping method, system, equipment and product

Abstract

This application discloses an automated geographic information system (GIS) mapping method, system, device, and product, relating to the fields of artificial intelligence and military information systems. The method includes: receiving natural language operational scenario text and performing secure de-identification preprocessing; inputting the de-identified text into a locally deployed, military-knowledge-enhanced inference model for semantic parsing, outputting structured data objects; generating precise geographic coordinates based on the location descriptions in the objects by querying a local military place name knowledge graph and coordinates, and backfilling the data; automatically verifying the coordinate-bearing data; and finally, automatically generating and mapping corresponding military symbols in the GIS according to the verification results and user permissions, obtaining a standardized military mapping layer. This application achieves end-to-end fully automated, high-precision conversion from unstructured text to standardized battlefield situation maps, significantly improving the efficiency and accuracy of military mapping while ensuring information security.

Description

Technical Field

[0001] This application relates to the fields of artificial intelligence and geographic information system technology, and in particular to an automated geographic information system mapping method, system, device and product. Background Technology

[0002] In the process of military operational scenario design and command decision-making, the rapid and accurate conversion of operational scenarios described in natural language (usually presented in document form) into standardized military plots on a geographic information system is a crucial link in supporting situational understanding and scenario simulation. Currently, this conversion process mainly relies on manual operation by command or plotting personnel, or is supplemented by semi-automated tools based on fixed rules and pre-made templates. This traditional model has inherent defects and has become a significant bottleneck restricting the efficiency and quality of operational planning.

[0003] The core problem facing existing technologies lies in the lack of automated and intelligent semantic understanding and parsing of natural language operational scenario texts, which contain rich military terminology, complex descriptions of geographical entities, and dynamic spatiotemporal relationships. Operational scenario texts are typical unstructured data objects, containing command intentions, troop formations, action sequences, and battlefield environment information that cannot be effectively identified and extracted by traditional keyword-matching or simple rule-based programs. For example, a system cannot accurately understand the entity type of Brigade A (our forces), the geographical reference of Hill 203, the relative spatial relationship of 5 kilometers to the east, and the time point H from a statement such as "Brigade A organizes defense 5 kilometers east of Hill 203 at time H," let alone correlate these elements into a complete tactical action entry.

[0004] Because the fundamental technical hurdle of natural language understanding could not be overcome, subsequent automated mapping processes could not be initiated. This directly led to two major consequences: First, the entire mapping process relied heavily on human reading, understanding, and manual operation, which was inefficient and prone to errors, making it difficult to meet the fast-paced decision-making needs of modern warfare; second, the effectiveness of any subsequent geographic information processing technologies (such as coordinate transformation and symbol rendering) was limited by this initial bottleneck, and even if they were advanced enough, they could not achieve end-to-end automation from text to map.

[0005] While some attempts exist to utilize general natural language processing techniques or publicly available geocoding services, these solutions generally suffer from a lack of military expertise, low recognition rates of military entities and tactical representations, and insufficient coordinate resolution accuracy. Furthermore, due to security requirements, military operational scenario texts typically contain classified information, and directly using general internet-based models or online services poses unacceptable security risks.

[0006] Therefore, the following technical problems urgently need to be solved in the existing technology: (1) the problem of automated semantic understanding of natural language combat scenarios; (2) the problem of high-precision and secure analysis of military geographic information; and (3) the problem of deployment security of large model capabilities under the premise of confidentiality. Summary of the Invention

[0007] This application aims to provide an automated geographic information system mapping method, system, device, and product, which can solve the technical problems that existing military scenario mapping methods rely on manual labor and are inefficient, and that existing automated solutions are difficult to put into practical use due to their inability to deeply understand military natural language, lack of professional geographic information analysis capabilities, and security risks.

[0008] To achieve the above objectives, this application provides the following solution: Firstly, this application provides an automated geographic information system mapping method, including: Receive natural language operational scenario text containing military semantics; perform security preprocessing on the operational scenario text to obtain de-identified structured paragraph units; The desensitized structured paragraph units are input into a locally deployed inference model enhanced with military domain knowledge for semantic understanding and parsing, and the output is a structured data object containing a location description field. Based on the location description in the structured data object, the corresponding geographic coordinates are parsed and generated, and the geographic coordinates are backfilled into the structured data object to generate a complete structured data object with coordinates; Perform output validation on the complete structured data object with coordinates; Based on the validated structured data objects and their contained geographic coordinates, the corresponding military symbols are automatically generated and plotted in the geographic information system to obtain a standardized military plotting layer.

[0009] Secondly, this application provides an automated geographic information system mapping system, comprising: The text receiving and preprocessing module is used to receive natural language operational scenario text containing military semantics and perform secure preprocessing to obtain de-identified structured paragraph units. The semantic understanding and parsing module is used to input the desensitized structured paragraph units into a locally deployed inference model enhanced with military domain knowledge, perform semantic understanding and parsing, and output a structured data object containing a location description field. The coordinate parsing and backfilling module is used to parse and generate corresponding geographic coordinates based on the location description in the structured data object, and backfill the geographic coordinates into the structured data object to generate a complete structured data object with coordinates; The output verification module is used to perform output verification on the complete structured data object with coordinates. The symbol generation and plotting module is used to automatically generate and plot corresponding military symbols in the geographic information system based on the validated structured data objects and their contained geographic coordinates, thereby obtaining a standardized military plotting layer.

[0010] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the automated geographic information system mapping method described in any one of the above-mentioned methods.

[0011] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the automated geographic information system mapping method described in any one of the above-mentioned methods.

[0012] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides an automated geographic information system (GIS) plotting method, system, device, medium, and product. By receiving natural language operational scenario text and performing secure preprocessing, it solves the problems of low efficiency and direct exposure of classified information in manual input, achieving automated triggering of plotting tasks and secure standardization of input data. By inputting desensitized structured paragraph units into a locally deployed inference model enhanced with military domain knowledge for semantic understanding and parsing, it solves the problems of general models being unable to understand military terminology, lacking domain knowledge, and having potential security leakage risks, achieving intelligent and structured information extraction from operational scenario text in a secure and reliable environment. By automatically generating and backfilling geographic coordinates based on the parsed location description, it solves the problems of insufficient coverage of military entities and low coordinate transformation accuracy in traditional geocoding services, achieving professional parsing from fuzzy text descriptions to precise military spatial data. By performing automated verification on complete structured data objects with coordinates and generating military symbols in the GIS accordingly, it solves the problems of low efficiency, error-proneness, and difficulty in ensuring standardization in manual plotting, achieving end-to-end fully automated and high-accuracy plotting from unstructured text to standardized battlefield situation maps. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the 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.

[0014] Figure 1 This is an application environment map of an automated geographic information system plotting method according to an embodiment of this application; Figure 2 A flowchart illustrating an automated geographic information system mapping method provided in one embodiment of this application; Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0017] The automated geographic information system mapping method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be set up independently, integrated into server 104, or placed in the cloud or on other servers. Terminal 102 can send the operational scenario text to be processed to server 104. After receiving the operational scenario text, server 104, based on a locally deployed reasoning model enhanced with military domain knowledge, automatically executes the complete process of text security preprocessing, semantic understanding and parsing, geographic coordinate generation and backfilling, output verification, and military symbol plotting, generating a standardized military plotting layer. Server 104 can provide feedback on the generated standardized military plotting layer or processing status to terminal 102. Furthermore, in some embodiments, an automated geographic information system plotting method can also be implemented independently by server 104 or terminal 102. For example, terminal 102 with corresponding computing power can directly perform localized plotting processing on the operational scenario text to be processed, or server 104 can obtain the operational scenario text to be processed from the data storage system and independently complete the automated plotting processing.

[0018] The terminal 102 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, portable wearable devices, and dedicated military command and information terminals, tactical tablets, etc. The server 104 can be implemented using independent servers or server clusters deployed in the military intranet or physically isolated environment, or it can be a private cloud server that meets the confidentiality requirements.

[0019] In one exemplary embodiment, such as Figure 2 As shown, an automated geographic information system mapping method is provided. This method is executed by computer equipment, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is described using a server as an example, and includes the following steps 1 to 5. Wherein: Step 1: Receive natural language operational scenario text containing military semantics; perform security preprocessing on the operational scenario text to obtain de-identified structured paragraph units; Step 2: Input the desensitized structured paragraph unit into a locally deployed inference model enhanced with military domain knowledge, perform semantic understanding and parsing, and output a structured data object containing a location description field; Step 3: Based on the location description in the structured data object, parse and generate the corresponding geographic coordinates, and backfill the geographic coordinates into the structured data object to generate a complete structured data object with coordinates; Step 4: Perform output validation on the complete structured data object with coordinates; Step 5: Based on the verified structured data objects and their contained geographic coordinates, automatically generate and plot the corresponding military symbols in the geographic information system to obtain a standardized military plotting layer.

[0020] By implementing steps 1 to 5 above, through receiving natural language operational scenario text, using a locally deployed military domain knowledge-enhanced inference model for semantic understanding and parsing, automatically extracting structured data objects, generating geographic coordinates, and automatically plotting military symbols, automated and intelligent understanding of natural language operational scenario text containing complex military terminology and spatiotemporal relationships is achieved, overcoming the fundamental bottleneck of traditional methods relying on manual parsing. The locally deployed military domain knowledge-enhanced inference model provides professional military semantic parsing capabilities while ensuring information security, laying a core foundation for the conversion from unstructured text to structured geographic information. End-to-end automated generation from natural language description to military symbol plotting is completed, significantly improving the overall efficiency and accuracy of operational scenario plotting.

[0021] Furthermore, this application can ensure that the entire processing complies with military confidentiality requirements and provide a structured data object foundation for subsequent intelligent decision support.

[0022] In another exemplary embodiment of this application, step 1 above, "receiving natural language operational scenario text containing military semantics; performing security preprocessing on the operational scenario text to obtain desensitized structured paragraph units," can be specifically implemented by steps 1.1 to 1.4 to constitute a complete front-end preprocessing flow, which securely and efficiently transforms the original, heterogeneous, and classified operational scenario text into unified, structured, and fully desensitized paragraph units, laying a reliable data foundation for the entire automated plotting process. Step 1.1: Receive the input raw combat scenario text, parse its document structure, and extract the plain text containing military semantics; Specifically: Receive electronic documents of operational scenarios in Word or PDF format uploaded from the command information system, call the corresponding format parsing library, remove all format control characters and non-text elements, and obtain complete, continuous plain text content.

[0023] Step 1.2: Based on the pre-set military classification dictionary and regular expression rules, the extracted plain text content is scanned in multiple rounds to automatically identify classified entities and automatically label the classified entities with classification levels according to the "Military Secrecy Regulations". Among them, classified entities include, but are not limited to, real geographical names, military unit designations, command organization names, equipment models, and names of important facilities.

[0024] The military classification dictionary contains sensitive entries organized by category (such as unit designations, command structures, equipment models, and important facilities), with each entry associated with a base classification level (public, secret, confidential, or top secret).

[0025] Regular expression rule sets are used to identify structured sensitive information, such as coordinate strings with specific formats (e.g., X point Y minute), communication call signs, time-encrypted codes, etc.

[0026] Multi-round scanning strategies include: The first round: Based on a military classification dictionary, precise matching and part-of-speech tagging are performed to identify explicitly classified entities (such as "123rd Motorized Infantry Brigade" and "Hill 203"). The second round: Apply regular expression rules to identify patterned sensitive information (e.g., "coordinates: confidential" "at H+1"). The third round: combining the identified classified entities with the context to identify implicit references (for example, when "my brigade" appears in the text, combine it with the previously identified unit number to determine its specific reference relationship).

[0027] Step 1.3: Using a reversible desensitization mechanism, the classified entities identified in Step 1.2 are replaced with uniform formatted codes in the corresponding positions of the plain text to obtain desensitized text; at the same time, a desensitization mapping table is generated and encrypted for storage.

[0028] The formatted token generation rule is as follows: it is generated using a uniform prefix "token_" followed by a sequence number (e.g., "token_001"). To further mitigate the possibility of the sequence number revealing the number of entities, the sequence number can be replaced with a random hash fragment (e.g., "token_a3f7").

[0029] The content of the desensitization mapping table: records the bidirectional mapping relationship between each format code and its corresponding original classified entity text, and includes entity type, security level label and operation information (such as operation timestamp, operator ID).

[0030] It should be noted that the desensitization mechanism here has a dual function: First, secure isolation between transmission and storage: The de-identified text does not contain any original confidential entities. Even if the intermediate results of subsequent processing are accessed without authorization, attackers can only obtain data containing codes and cannot restore the original information.

[0031] Second, semantic preservation in inference: By using Few-shot hints containing examples of desensitized codes, subsequent inference models can understand that the code "token_XXX" is a placeholder for a specific entity. Thus, without touching the original classified content, it can still accurately complete entity type judgment (e.g., judging "token_001" as "military unit type") and semantic relationship parsing (e.g. parsing the spatial semantics of "token_002 30 kilometers east").

[0032] Step 1.4: Based on the five elements of military professional elements—mission, enemy situation, our own situation, terrain, and time—the desensitized text is divided and categorized into paragraphs, and the desensitized structured paragraph units are output.

[0033] By executing steps 1.1 to 1.4 above, the automatic parsing and plain text extraction of original documents (original operational scenario texts) in different formats were first achieved, eliminating manual intervention and providing a standardized input starting point. Secondly, classified information was automatically identified and labeled based on rules, implementing confidentiality requirements from the source. Thirdly, a reversible desensitization mechanism enabled core processing to be conducted in a secure environment, while ensuring the recoverability of authorized information. Finally, the text was structurally segmented based on military elements, optimizing the input quality and processing efficiency of subsequent semantic parsing.

[0034] In another exemplary embodiment of this application, step 2 above, "inputting the desensitized structured paragraph unit into a locally deployed inference model enhanced with military domain knowledge, performing semantic understanding and parsing, and outputting a structured data object containing a location description field," can be specifically implemented by steps 2.1 to 2.5 to achieve accurate conversion from natural language text to machine-readable structured combat elements: Step 2.1: Receive the desensitized structured paragraph units output from Step 1, and assemble them with the preset prompt template to form prompt text that conforms to the model input format; Step 2.2: In a physically isolated local server environment, load the publicly available base language model; the weight file of the base language model is stored and loaded in plaintext. Step 2.3: Decrypt and dynamically load the military-specific model fine-tuning adapter from the encrypted storage medium, and attach the adapter to the loaded base large language model to form a reasoning model with military-domain knowledge enhancement capabilities; Furthermore, the aforementioned model fine-tuning adapter is obtained by fine-tuning the base large language model using Low-Rank Adaptive (LoRA) technology. It is generated through training with a limited number of high-quality military-labeled samples and only updates a portion of the weight matrix in the attention mechanism of the base large language model. By attaching the LoRA adapter (i.e., the model fine-tuning adapter) to the loaded, frozen base large language model, a reasoning model with military domain knowledge enhancement capabilities is dynamically formed without changing the original parameters of the base large language model. The training dataset for the military-specific LoRA adapter consists of the following three types of samples: The first category consists of military entity identification samples, including troop unit designations, tactical elements (high ground, crossings, bridges, etc.), and weapon and equipment models, totaling approximately 200 samples. The second category consists of relative position description analysis samples, which include paired samples of position descriptions in the format of "reference point + azimuth + distance" and corresponding structured analysis results, totaling approximately 180 samples. The third category consists of military symbol code mapping samples, which include approximately 120 samples mapping tactical action description texts to the corresponding national military standard GJB1839A-2019 symbol codes.

[0035] All the above samples were annotated by professionals with relevant expertise and underwent cross-validation, achieving an annotation consistency rate of no less than 92%. Using this dataset, a large language model was built upon with a learning rate of 2×10⁻⁶. -4 After 3 training epochs of fine-tuning, the F1 score for military entity recognition reached 0.89 on the validation set, and the location description parsing was also improved.

[0036] In a specific implementation, Qwen-14B or Llama2-13B is used as the publicly available pedestal large language model. The LoRA adapter is configured with a rank r=16 and a scaling factor α=32, and is trained using approximately 500 high-quality military-labeled samples. The trained LoRA adapter file is approximately 16MB in size. For security, the LoRA adapter file is encrypted using the national cryptographic algorithm SM4 and dynamically decrypted and mounted during inference. The model's inference process can be expressed as h = W0x + BAx, where W0 is the frozen pedestal model weights, and BA is the low-rank increment introduced by the LoRA adapter.

[0037] Step 2.4: Input the assembled prompt text into the inference model. The inference model performs semantic understanding and parsing inference on the input prompt text and outputs a structured data object. Specifically, the structured data object (e.g., a JSON object) contains multiple fields parsed from the text, mainly including: entity type (entity_type), unit name (unit_name, with a de-identified code, such as "token_001"), location description (location_desc, which may contain a de-identified code, such as "token_002 30 km east"), timestamp (timestamp, such as "D day H hour"), and military symbol code (symbol_code, such as "SFGPCICV---"). Additionally, the object contains a reserved geographic coordinate field (coordinate), initially empty (null), for later backfilling.

[0038] Step 2.5: After the inference is completed, unload the model fine-tuning adapter from the inference model and clear the intermediate data and computational traces of this inference task in memory.

[0039] Specifically, after inference is complete, the adapter management interface is called to perform the following operations: unload the LoRA adapter weight matrix from the GPU memory and release the reference count of the corresponding memory page; perform an explicit zeroing operation (memset to zero) on the memory region storing intermediate activation values ​​and then release it; record the operation log after the above operations are completed. It should be noted that the above operations can effectively prevent subsequent access within the same process, but cannot replace the data destruction procedures of storage media that comply with national security standards. If higher security requirements are required, it should be used in conjunction with hardware-level memory isolation mechanisms.

[0040] By executing steps 2.1 to 2.5 above, secure and efficient military semantic parsing was achieved. This process ensures input compatibility and the reliability of the inference model interface. Performance degradation is avoided by loading the base large language model in plaintext within an isolated environment, and secure and flexible injection of professional knowledge is achieved through dynamically attaching an encrypted dedicated adapter. The enhanced inference model can deeply understand military operational scenario texts and output accurate machine-readable structured data objects. Immediately after inference, the model fine-tuning adapter is unloaded and memory is cleaned, ensuring a high level of security isolation and effectively preventing the risk of high-secret knowledge residue and data leakage.

[0041] In another exemplary embodiment of this application, step 3 above, "based on the location description in the structured data object, parse and generate the corresponding geographic coordinates, and backfill the geographic coordinates into the structured data object to generate a complete structured data object with coordinates," can be specifically implemented by steps 3.1 to 3.3 to complete the parsing and generation from text location description to precise geographic coordinates: Step 3.1: Receive the structured data object output from Step 2, parse the content of the position description field in the structured data object, and determine its type; Furthermore, the types of location descriptions include absolute location descriptions and relative location descriptions; among them, absolute location descriptions include "Hill 203"; relative location descriptions include "5 kilometers east of Hill 203".

[0042] Step 3.2: Based on the type of the location description field, adopt the corresponding coordinate acquisition and calculation strategy: If it is an absolute location description: directly query the local military place name knowledge graph to obtain the coordinates of the named entity as preliminary coordinates.

[0043] If the description is a relative position, then perform the following sub-steps: Parse the reference point (e.g., "Hill 203"), azimuth (e.g., "East" corresponds to 90°) and distance (e.g., "5 km") from the location description field; Query the local military place name knowledge graph to obtain the coordinates of the reference point; Based on the analyzed distance parameters, an adaptive geodesy calculation model is selected to obtain the preliminary coordinates of the target point.

[0044] The above adaptive algorithm strategy is as follows: When the distance d ≤ 30 km, a planar approximation model is used for calculation. Preliminary coordinates (longitude) are calculated. with latitude : ; ; ; ; in, and These are the latitude and longitude offsets of the target point relative to the reference point (unit: degrees); d is the distance (in meters), and θ is the azimuth (in degrees, measured clockwise from due north). The average radius of the Earth (unit: meters); and The coordinates are the latitude and longitude (in radians) of the reference point obtained from the local military place name knowledge graph. Under these conditions, the additional error introduced by the planar approximation does not exceed 0.15 meters, which meets the accuracy requirements.

[0045] When the distance is 30 km < d ≤ 500 km, the Vincenty iterative algorithm is used to calculate the coordinates of the target point. The convergence accuracy of this algorithm on the WGS84 ellipsoid is better than 0.1 mm, and the error introduced by the algorithm itself can be ignored.

[0046] When the distance d > 500 kilometers, the system automatically marks the entry as pending_review, prompting manual confirmation. This is because describing relative positions at extremely long distances is considered an abnormal situation in tactical-level operational scenarios.

[0047] The local military place name knowledge graph contains approximately 8,000 military entities, of which 5,500 are tactical elements (such as hills, bridges, and crossings).

[0048] Step 3.3: Standardize the preliminary coordinates obtained or calculated in Step 3.2: The preliminary coordinates obtained or calculated in step 3.2 are then uniformly converted to the target coordinate system (such as CGCS2000) required by the geographic information system platform using a packaged coordinate transformation service (such as the Proj4 library). This system supports high-precision conversion between coordinate systems such as WGS84, CGCS2000, UTM, and MGRS.

[0049] By calling digital elevation model data, the elevation of the transformed plane coordinates is estimated and corrected to obtain the final target point coordinates including longitude, latitude, and elevation.

[0050] The target point coordinates are then filled back into the geographic coordinates field of the structured data object, and a complete structured data object with coordinates is output.

[0051] Through the above coordinate transformation and elevation correction processes, the horizontal position error can be controlled within 0.3 meters, and the elevation error within 0.5 meters. The geographic information system can use an error propagation model to evaluate the overall coordinate accuracy, and its total error... The estimation formula is: ; in, For datum transformation error (e.g., ); This refers to map projection distortion error; For the Military Grid Reference System (MGRS), the grid truncation error is related to the grid resolution. To verify the accuracy of the above coordinate analysis and transformation process, an evaluation was conducted on a test set of 50 military place names containing known coordinates. The test samples covered four types of tactical elements: high ground, ferry crossings, bridges, and road intersections, with an absolute position description to relative position description sample ratio of approximately 3:2. Test results show that for absolute position descriptions, the mean square error of the planar position obtained by direct querying from the knowledge graph is 0.28 meters; for relative position descriptions, when the reference distance is no more than 30 kilometers, the mean square error of the planar position calculated using the planar approximation model is 0.31 meters; when the reference distance exceeds 30 kilometers, using the Vincenty algorithm, the mean square error of the planar position can be controlled within 0.35 meters. The mean square error of the elevation after elevation correction is 0.47 meters.

[0052] It should be noted that the accuracy of the above-mentioned planar approximation model will decrease significantly when the reference distance exceeds 100 kilometers. In this scenario, a strict geodetic algorithm should be forcibly enabled.

[0053] By executing steps 3.1 to 3.3 above, the location descriptions in the structured data objects are first automatically identified and classified, providing a clear basis for solving differentiated coordinates. Secondly, by prioritizing queries of the local military place name knowledge graph, the problem of insufficient coverage of military entities by general geographic services is solved, ensuring the professionalism and security of coordinate acquisition. Furthermore, adaptively selected geodetic formulas are applied for relative position calculation, and high-precision coordinate system transformation and elevation correction are performed, converting ambiguous text descriptions into accurate geographic coordinates that meet military application requirements, overcoming the problems of large errors in simple conversions and inconsistent coordinate benchmarks. Finally, coordinate backfilling achieves a closed loop in the data flow, providing an accurate spatial data foundation for subsequent plotting.

[0054] In another exemplary embodiment of this application, step 4, "output validation of the structured data object with coordinates," can be specifically implemented through steps 4.1 to 4.5 to achieve automated rule validation, quality filtering, and anomaly security handling. Step 4.1: Receive the complete structured data object with coordinates from the output of Step 3, and load the predefined integrity verification rule base; Step 4.2: Perform automated validation on the coordinates contained in the geographic coordinates field of the complete structured data object, and output the validation results, specifically including: C1) Check whether the coordinates contained in the geographic coordinates field are within the preset legal war zone, training area, or mission range; C2) Check whether the coordinates contained in the geographic coordinates field violate geographic constraints; for example, if they fall within national borders, sensitive areas, large bodies of water, or are obviously inconsistent with the entity type (e.g., marking the coordinates of a ship on land). C3) Combine entity type, military symbol code and terrain attributes obtained from local military place name knowledge graph to make a reasonable judgment (for example, check whether the coordinates of the helicopter landing site are located in a relatively flat area; whether the military symbol code conforms to the national military standard "GJB 1839A-2019"). Step 4.3: Update the status flag of the complete structured data object based on the verification results: If the validation passes completely, the status of the complete structured data object is marked as valid (or pending rendering). If the verification fails, the complete structured data object that requires manual review will be marked as review, the data that is pending because the reference point was not found will be marked as pending_review, and the data that contains fundamental errors will be marked as invalid. The specific reasons for the anomaly will be recorded in the data (such as coordinates exceeding the task area, terrain and entity type conflict). Step 4.4: Based on the status flags, add the complete structured data objects marked as valid directly to the rendering queue and execute step 5; import the complete structured data objects marked as review into the manual review database and trigger an alarm to notify authorized personnel to perform manual review. Step 4.5: After manual review, the corrected coordinates are resubmitted to Step 4.1 for verification. If the process is stuck due to fundamental errors such as the inability to resolve the target point coordinates, the exception will be captured, the status will be marked as error, the log will be recorded, and the data will be skipped to ensure that the processing of other data is not blocked.

[0055] Steps 4.1 to 4.5 above together constitute an automated, rule-driven output verification and quality control process. This process loads a military geography professional knowledge rule base to verify the compliance, rationality, and consistency of the generated geographic coordinates, and automatically completes data distribution (rendering, review, and archiving) based on the verification results. This not only intercepts non-compliant data caused by model illusions or parsing errors at the source, ensuring the reliability of data entering the plotting stage, but also ensures the smoothness and closed loop of the entire process through manual review and anomaly handling mechanisms, providing high-quality data filtering and security for automated plotting.

[0056] In another exemplary embodiment of this application, step 5 above, "automatically generating and plotting corresponding military symbols in the geographic information system based on the verified structured data object and its contained geographic coordinates, and obtaining a standardized military plotting layer," can be specifically implemented by steps 5.1 to 5.6 to achieve automatic, secure, and adaptive conversion from structured data to standardized visual military symbols: Step 5.1: Receive the complete structured data object with coordinates from the rendering queue in Step 4; based on the entity type and military symbol code field in the object, match the corresponding Scalable Vector Graphics template from a pre-built standardized military symbol library (e.g., an SVG template library containing approximately 1200 point, line, and polygon military symbols); instantiate the Scalable Vector Graphics (SVG) template into a spatial graphic object carrying the geographic coordinate field; Step 5.2: Based on the current operator's identity and permission level, query the desensitization mapping table generated in Step 1.3, and selectively restore the classified entity text in the structured data associated with the spatial graphic object to determine its final text labeling content on the map.

[0057] The selective restoration operation includes: receiving the current operator's permission level, a de-identification mapping table, and structured data containing a de-identification code (such as "token_001"); querying the mapping table according to the permission level to determine whether the operator has the right to view the original text; and executing the corresponding restoration strategy (such as complete restoration, partial de-identification, or rejection of display) based on the determination result.

[0058] Step 5.3: Obtain the map scale denominator of the current geographic information system view. Based on this scale denominator, dynamically calculate the display scaling factor of spatial graphic objects using a predefined formula, and adjust the size and line width of the spatial graphic objects in real time using attributes such as transform="scale(k)" to ensure that they are clearly distinguishable and conform to cartographic specifications at different scaling levels; Specifically, based on the current scale denominator S, the display scaling factor k of spatial graphic objects is dynamically calculated using a predefined formula: in, The denominator of the base scale, for example =50000; Step 5.4: Call the application programming interface (API) provided by the geographic information system platform to add the configured spatial graphic objects in batches to the dedicated symbol layer, and overlay and render the symbol layer onto the corresponding coordinate positions on the electronic map; Step 5.5: Monitor the total number of spatial graphic objects to be rendered within the current view range in real time; when the total number exceeds the preset performance threshold, automatically trigger the detail level optimization mechanism to aggregate and draw multiple adjacent or similar spatial graphic objects into a composite military symbol; Step 5.6: Render and overlay all spatial graphic objects configured and optimized in the above steps onto the electronic map through the geographic information system platform to generate a complete standardized military plotting layer.

[0059] By implementing steps 5.1 to 5.6 above, this application achieves automated, secure, and controllable generation of structured data with coordinates into standardized military plotting layers. This process automatically matches and instantiates scalable vector graphics templates based on standardized codes, completing a direct and accurate conversion from data to standardized graphics, eliminating human error and delays, and ensuring the standardization and consistency of the output results. By dynamically querying the anonymized mapping table based on user permissions and selectively restoring classified information, the system achieves hierarchical and on-demand visualization of classified information, effectively reconciling the conflict between data confidentiality and command integrity while ensuring information security. Simultaneously, the system dynamically adjusts the display parameters of symbols according to the current map scale, ensuring good visual recognizability and cartographic standardization at different scaling levels, overcoming the problem of fixed-size symbols being unreadable or having a chaotic layout when scaling. By calling the geographic information system interface to automatically complete layer creation and map overlay, the system achieves the final rendering from geographic coordinates to screen coordinates, transforming all processing results into an intuitive battlefield situation map. Furthermore, by monitoring the number of elements to be rendered in real time and automatically triggering a detail level optimization mechanism, system interface lag caused by an excessive number of elements is effectively prevented, ensuring smooth interaction and user experience during large-scale plotting. These steps together constitute a complete plotting generation pipeline, ultimately producing a standardized military plotting layer that can be directly used for operational analysis.

[0060] As an optional implementation, the method further includes task management and optimization steps: It receives and manages multiple concurrent operational scenario texts, and performs intelligent scheduling and resource allocation based on the priority of the different operational scenario texts and their estimated computing resource requirements.

[0061] The structured data object with coordinates generated in step 3 is cached. When a new combat scenario text is received, if it is determined that it is consistent with the core entity, location and action description of the cached mission, the cached structured data object is called first and the subsequent verification and plotting process is directly entered to avoid repeated semantic parsing and coordinate calculation.

[0062] As an optional implementation, the method further includes steps for knowledge base maintenance and optimization: A knowledge graph management interface is provided, supporting authorized users to securely review, add, modify, or disable entities, coordinates, and related attributes in the local military place name knowledge graph. Cases where manual review is triggered during the verification process in step 4 due to the absence of corresponding entities in the local military place name knowledge graph are recorded. Valid geographic coordinates and their descriptive text, after manual correction and confirmation, can be securely reviewed and added to the knowledge graph through the management interface, thereby achieving continuous optimization of geographic information parsing capabilities.

[0063] As an optional implementation, the method further includes a security audit and monitoring step: Generate and protect operation audit logs: After each key operation node from step 1 to step 5 of the method, an immutable audit log is automatically generated, recording the operation content, result hash value, executor identification and timestamp, and stored in encrypted form using national cryptographic algorithms.

[0064] Monitoring and early warning of abnormal operation modes: Based on predefined rules or models, real-time monitoring and alarms are performed on abnormal modes in the processing flow; the abnormal modes include, but are not limited to: multiple consecutive coordinate parsing failures, and high-frequency access to the original information corresponding to a specific de-identified entity within a short period of time.

[0065] This application also provides an application scenario in which the aforementioned automated geographic information system (GIS) plotting method is applied. The automated GIS plotting method described in this application can be widely used in scenarios such as combat planning and simulation exercises. The following explanation uses a brigade-level red-blue force-on-force tactical scenario plotting as an example.

[0066] Traditional method: Staff officers manually convert a combat scenario document of about 500 words (《Red Force Motorized Infantry Brigade Attacking Blue Force Heights Exercise Scenario.docx》) into an electronic situation map, which usually takes 2-3 hours and is prone to errors. Using the method of this application: End-to-end automated processing via a geographic information system, the entire process takes about 4 minutes. The specific process follows the steps described in the embodiments of this application: Step 1: Preprocessing and Desensitization: The Geographic Information System automatically identifies and desensitizes classified entities in the operational scenario document (e.g., "123rd Motorized Infantry Brigade" is desensitized to "token_001"), and generates an encrypted desensitization mapping table.

[0067] Step 2, Semantic parsing: Input the anonymized text into the local inference model and output structured data, such as: {Entity type: "Mobulv", Unit: "token_001", Location: "token_002", ...}.

[0068] Step 3: Coordinate Generation: Query the local military place name knowledge graph, parse "token_002" (corresponding to Hill 203) into precise geographic coordinates [118.652338, 32.123452,352] and backfill them into the structured data.

[0069] Step 4, Output Verification: Automatically verify whether the coordinates are located in a legal training area, whether the symbols conform to national military standards, and whether the terrain matches. Verification passes.

[0070] Step 5: Symbol plotting: Based on permissions, restore the display to "123rd Motorized Infantry Brigade". At a map scale of 1:50000, automatically render the corresponding military symbols to the coordinate positions and generate a standardized military plotting layer.

[0071] Based on the same inventive concept, this application also provides an automated geographic information system (GIS) mapping system for implementing the automated GIS mapping method described above. The solution provided by this system is similar to the implementation scheme described in the above method. Therefore, the specific limitations of one or more embodiments of the automated GIS mapping system provided below can be found in the limitations of the automated GIS mapping method described above, and will not be repeated here.

[0072] In one exemplary embodiment, an automated geographic information system mapping system is provided, comprising: The text receiving and preprocessing module is used to receive natural language operational scenario text containing military semantics and perform secure preprocessing to obtain de-identified structured paragraph units. The semantic understanding and parsing module is used to input the desensitized structured paragraph units into a locally deployed inference model enhanced with military domain knowledge, perform semantic understanding and parsing, and output a structured data object containing a location description field. The coordinate parsing and backfilling module is used to parse and generate corresponding geographic coordinates based on the location description in the structured data object, and backfill the geographic coordinates into the structured data object to generate a complete structured data object with coordinates; The output verification module is used to perform output verification on the complete structured data object with coordinates. The symbol generation and plotting module is used to automatically generate and plot corresponding military symbols in the geographic information system based on the validated structured data objects and their contained geographic coordinates, thereby obtaining a standardized military plotting layer.

[0073] As an optional implementation, the text receiving and preprocessing module includes a security classification scanning unit, a reversible desensitization unit, and a paragraph segmentation unit. The security classification scanning unit, based on a preset military security classification dictionary and regular expression rules, performs multiple scans on the input original operational scenario text, automatically identifying classified entities and labeling them with their security classifications, outputting a list of entities with security classification labels. The reversible desensitization unit, based on the entity list, replaces the identified classified entities in the corresponding positions of the plain text with a formatted code consisting of a unified prefix and a sequence number (or a random hash), generating desensitized text and an encrypted desensitization mapping table. The paragraph segmentation unit divides and categorizes the desensitized text into paragraphs according to the five elements of the military scenario: "mission, enemy situation, friendly situation, terrain, and time," outputting structured desensitized paragraph units.

[0074] As an optional implementation, the semantic understanding and parsing module includes a model loading unit, an inference execution unit, and a model unloading unit. The model loading unit is responsible for decrypting and dynamically mounting a military-specific adapter, fine-tuned using Low-Rank Adaptive (LoRA) technology, from a storage medium encrypted with national cryptographic algorithms onto the loaded base large language model, forming an inference model with enhanced military domain knowledge. The inference execution unit receives the structured paragraph unit, assembles it with a preset few-sample prompt template, inputs it into the inference model, performs semantic understanding and parsing, and outputs a structured data object containing entity type, unit name (de-identified code), location description (possibly containing a de-identified code), timestamp, military symbol code, and reserved null value geographic coordinate fields. After inference is complete, the model unloading unit immediately unloads the LoRA adapter from the base model and explicitly zeroes out intermediate data and computational traces in memory to ensure no information remains.

[0075] As an optional implementation, the coordinate parsing and backfilling module includes a location type judgment unit, a knowledge graph query unit, a geodetic coordinate calculation unit, and a coordinate backfilling unit. Each unit processes the data sequentially and writes the final coordinates into the reserved fields of the structured data object. Specifically, the location type judgment unit parses the location description field in the structured data object to determine whether it is an absolute or relative location description; the knowledge graph query unit directly queries the local military place name knowledge graph to obtain preliminary coordinates for absolute location descriptions, and first parses the reference point name for relative location descriptions, then queries the knowledge graph to obtain the reference point coordinates; the geodetic coordinate calculation unit adaptively selects a calculation model based on the parsed distance parameters: a planar approximation model is used when the distance is ≤30 km, and the Vincenty iterative algorithm is used when the distance is ≤500 km, to calculate the preliminary coordinates of the target point; the coordinate backfilling unit converts the preliminary coordinates to the target coordinate system using a coordinate transformation service, calls a digital elevation model for elevation correction, obtains the final geographic coordinates, and then backfills them into the reserved coordinate fields of the structured data object.

[0076] As an optional implementation, the output verification module includes a rule base loading unit, a multi-dimensional verification unit, and a state distribution unit. The rule base loading unit preloads geographical range rules, national military standard symbol specifications, and spatiotemporal logic rules. The multi-dimensional verification unit sequentially performs range verification (whether it is located within a legal war zone), boundary and terrain verification (whether it falls outside national borders, waterways, or contradicts entity types), and rationality verification (combining entity type, symbol code, and terrain attributes for logical judgment). The state distribution unit marks the data object status as "valid" and sends it to the rendering queue based on the verification results, or marks it as "review" (or "pending_review", "invalid") and writes it to the manual review database, while recording the reason for the anomaly to ensure the main process is not blocked.

[0077] As an optional implementation, the symbol generation and plotting module includes an SVG template matching unit, an access control restoration unit, a scale adaptation unit, and a layer rendering unit. Data is transferred between these units via spatial graphic objects. Specifically, the SVG template matching unit matches a corresponding scalable vector graphic template from a pre-set standardized military symbol library based on the entity type and military symbol code in the validated structured data object, and instantiates it into a spatial graphic object carrying geographic coordinates. The access control restoration unit, based on the current operator's permissions, queries a desensitization mapping table to selectively restore the classified entity text associated with the spatial graphic object, determining the final text annotation content displayed on the map. The scale adaptation unit obtains the scale denominator of the current map view, dynamically calculates the scaling factor and line width of the symbols using a predefined formula, and adjusts the visual style of the spatial graphic objects accordingly. The layer rendering unit calls the application programming interface of the geographic information system platform to add the configured spatial graphic objects in batches to a dedicated symbol layer, overlays and renders them onto the electronic map, and automatically triggers a detail level optimization mechanism for aggregation display when the number of symbols exceeds a performance threshold, ultimately generating a standardized military plotting layer.

[0078] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 3 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media to run. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements an automated geographic information system mapping method.

[0079] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0080] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0081] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0082] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0083] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0084] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0085] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0086] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0087] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An automated geographic information system mapping method, characterized in that, Includes the following steps: Receive natural language operational scenario text containing military semantics, perform security preprocessing on the operational scenario text, and obtain desensitized structured paragraph units; The desensitized structured paragraph units are input into a locally deployed inference model enhanced with military domain knowledge for semantic understanding and parsing, and the output is a structured data object containing a location description field. Based on the location description in the structured data object, the corresponding geographic coordinates are parsed and generated, and the geographic coordinates are backfilled into the structured data object to generate a complete structured data object with coordinates; Perform output validation on the complete structured data object with coordinates; Based on the validated structured data objects and their contained geographic coordinates, the corresponding military symbols are automatically generated and plotted in the geographic information system to obtain a standardized military plotting layer.

2. The automated geographic information system mapping method according to claim 1, characterized in that, The process of performing security preprocessing on the operational scenario text to obtain desensitized structured paragraph units specifically includes: The document structure of the operational scenario text is parsed to extract the plain text containing military semantics; Based on a pre-set military classification dictionary and regular expression rules, the plain text is scanned to automatically identify classified entity texts and mark the classification level of the identified classified entity texts. A reversible desensitization mechanism is adopted to replace the identified classified entity text with a formatted code to obtain desensitized text, and an encrypted desensitization mapping table is generated that records the mapping relationship between the formatted code and the original classified entity text. The desensitized text is divided and categorized into paragraphs according to military professional elements, and the desensitized structured paragraph units are output.

3. The automated geographic information system mapping method according to claim 2, characterized in that: The formatted code is generated using a unified prefix plus a sequence number. The desensitization mapping table also records the entity type, security classification, and operation information of the classified entity text.

4. The automated geographic information system mapping method according to claim 1, characterized in that, The construction and execution of the inference model specifically includes: Load the publicly available base language model in a physically isolated local environment; Decrypt and dynamically load a military-grade model fine-tuning adapter from encrypted storage media; The model fine-tuning adapter is attached to the base large language model to form a reasoning model with enhanced military domain knowledge capabilities; The structured paragraph unit, which is assembled with the prompt template, is input into the reasoning model to obtain the structured data object; After inference is completed, the model fine-tuning adapter is unloaded from the inference model, and the intermediate data of this inference task in memory is cleared.

5. The automated geographic information system mapping method according to claim 1, characterized in that, The step of parsing and generating corresponding geographic coordinates based on the location description in the structured data object specifically includes: Determine whether the location description is an absolute location description or a relative location description; If it is an absolute location description, then query the local military place name knowledge graph to obtain the coordinates of the corresponding named entity; If it is a relative position description, the reference point, azimuth, and distance are parsed from the position description; the coordinates of the reference point are obtained by querying the local military place name knowledge graph; based on the coordinates, azimuth, and distance of the reference point, the preliminary coordinates of the target point are calculated using geodetic formulas.

6. The automated geographic information system mapping method according to claim 1, characterized in that, The output validation of the complete structured data object with coordinates specifically includes: Check whether the geographic coordinates are within a preset legal geographic range; check whether the geographic coordinates violate predefined geographic constraints; combine the entity type, military symbol code and terrain attributes in the structured data object to make a reasonableness judgment and obtain the verification result; Update the status flag of the complete structured data object based on the verification results, and import the object marked as requiring review into the manual review process.

7. The automated geographic information system mapping method according to claim 1, characterized in that, The automatic generation and plotting of corresponding military symbols to obtain a standardized military plotting layer specifically includes: Based on the entity type and military symbol code in the structured data object, match the corresponding scalable vector graphic template and instantiate it as a spatial graphic object carrying the geographic coordinates; Based on the current operator's permission level, query the desensitization mapping table to selectively restore the classified entity text associated with the spatial graphic object; Based on the map scale of the current geographic information system view, the display parameters of the spatial graphic objects are dynamically adjusted. The configured spatial graphic object is added to the symbol layer and overlaid and rendered onto the electronic map to obtain a standardized military plotting layer.

8. An automated geographic information system mapping system, characterized in that, include: The text receiving and preprocessing module is used to receive natural language operational scenario text containing military semantics and perform secure preprocessing to obtain de-identified structured paragraph units. The semantic understanding and parsing module is used to input the desensitized structured paragraph units into a locally deployed inference model enhanced with military domain knowledge, perform semantic understanding and parsing, and output a structured data object containing a location description field. The coordinate parsing and backfilling module is used to parse and generate corresponding geographic coordinates based on the location description in the structured data object, and backfill the geographic coordinates into the structured data object to generate a complete structured data object with coordinates; The output verification module is used to perform output verification on the complete structured data object with coordinates. The symbol generation and plotting module is used to automatically generate and plot corresponding military symbols in the geographic information system based on the validated structured data objects and their contained geographic coordinates, thereby obtaining a standardized military plotting layer.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of an automated geographic information system mapping method according to any one of claims 1-7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the automated geographic information system mapping method according to any one of claims 1-7.

Images