Automatic quality inspection method, system, device and medium based on low utility land data

Abstract

The application relates to an automatic quality inspection method, system, device and medium based on low-utility land data, and belongs to the technical field of land use monitoring. The automatic quality inspection method comprises the following steps: receiving a low-utility land policy standard document input by a user; performing semantic analysis on the low-utility land policy standard document, identifying rule entities and constraint logic, and generating a structured latest standard rule set; performing difference comparison between the latest standard rule set and a pre-stored historical quality inspection rule library, generating a rule update instruction set according to the difference comparison result; dynamically updating the quality inspection rule library according to the rule update instruction set, and outputting a synchronously updated quality inspection rule library; based on the synchronously updated quality inspection rule library, performing automatic quality auditing on target low-utility land data to be processed, and generating an original quality inspection result set; analyzing defect distribution characteristics of the original quality inspection result set, and generating a customizable quality inspection report. The application can improve the quality inspection accuracy and timeliness of low-utility land data.

Description

Technical Field

[0001] This application relates to the field of land use monitoring technology, and in particular to an automatic quality inspection method, system, equipment and medium based on inefficient land use data. Background Technology

[0002] As my country's urbanization process continues to accelerate and land resources become increasingly scarce, local governments are gradually strengthening their efforts to identify, assess, and redevelop existing construction land, especially inefficient land. Against this backdrop, scientifically and rationally defining whether a plot of land constitutes "inefficient land" has become a crucial aspect of improving land use efficiency and optimizing the spatial layout of the national territory. Typically, local governments will issue relevant policy documents, such as the "Administrative Measures for the Redevelopment of Inefficient Industrial Land" and related implementing rules, as the basic basis for identifying inefficient land.

[0003] However, in current practice, these policy standards mostly exist in unstructured text form, commonly in PDF, Word, or TXT formats. The content is highly specialized and complex, lacking unified data expression standards and machine readability support. This leads to a significant need for manual intervention in practical applications, especially during large-scale land use data quality reviews. This is not only time-consuming and labor-intensive but also prone to biases due to subjective judgment, affecting the accuracy and consistency of the overall work. Furthermore, since inefficient land use standards are dynamically adjusted, policies issued in different regions, and even within the same region at different times, may differ, making it particularly difficult to maintain the timeliness and completeness of quality inspection rules. Especially when faced with massive amounts of data to be inspected, the traditional method of relying on manual experience to compare each item is clearly insufficient to meet the requirements of modern government information systems for efficient, accurate, and intelligent processing capabilities. Summary of the Invention

[0004] To improve the efficiency and accuracy of quality inspection of inefficient land use data, this application provides an automatic quality inspection method, system, equipment, and medium based on inefficient land use data.

[0005] Firstly, this application provides an automatic quality inspection method based on inefficient land use data, employing the following technical solution:

[0006] An automated quality inspection method based on inefficient land use data, the quality inspection method comprising:

[0007] Receive user input of inefficient land use policy standard documents;

[0008] Semantic parsing is performed on the aforementioned inefficient land use policy standard document to identify rule entities and constraint logic, and a structured latest standard rule set is generated;

[0009] The latest standard rule set is compared with the pre-stored historical quality inspection rule base, and a rule update instruction set is generated based on the comparison results.

[0010] The quality inspection rule base is dynamically updated according to the rule update instruction set, and the synchronously updated quality inspection rule base is output.

[0011] Based on the synchronously updated quality inspection rule base, automated quality audits are performed on the target inefficient land use data to be processed, generating an original quality inspection result set;

[0012] The defect distribution characteristics of the original quality inspection result set are analyzed to generate a customizable quality inspection report.

[0013] By adopting the above technical solutions, a closed-loop intelligent quality inspection system was constructed. Natural Language Processing (NLP) automatically parses inefficient land use policy standards and generates a structured rule set. Combined with a hash comparison mechanism, the quality inspection rule base is dynamically updated. A dual-mode rule engine (general algorithm + custom script) performs quality audits on standardized inefficient land use data. Finally, based on Natural Language Generation (NLG) technology, customizable graphic and textual reports are output. Compared to traditional methods relying on human intervention, this approach significantly improves the accuracy, timeliness, and sustainable development capabilities of quality inspection work, contributing to a comprehensive improvement in the level of refined natural resource management under the background of government digital transformation.

[0014] Optionally, the steps of semantically parsing the inefficient land use policy standard document, identifying rule entities and constraint logic, and generating a structured, up-to-date standard rule set include:

[0015] Extract text content elements from the inefficient land use policy standard document to generate an initial set of semantic units;

[0016] Entity recognition is performed on the initial semantic unit set, and field names, data table names, and constraint entities are labeled.

[0017] Based on a pre-trained language model, the logical relationships between labeled entities are parsed, the dependency relationships between entities are identified, and an enhanced semantic unit set is generated.

[0018] The set of enhanced semantic units is transformed into structured rule tuples, generating the latest standard rule set.

[0019] By employing the aforementioned technical solution, multi-level and multi-dimensional semantic parsing and logical modeling of standard documents related to inefficient land use policies are performed, achieving automatic conversion from unstructured policy text to structured rule sets. This solution not only effectively solves the problems of low efficiency and error-proneness in traditional manual rule entry but also enhances the understanding of complex semantic relationships by introducing a pre-trained language model, resulting in higher accuracy and generalization capabilities throughout the parsing process. Particularly in the field of land resource management, this method helps promote the digital implementation of policies and rules, providing solid data support and a rule foundation for application scenarios such as intelligent approval, compliance review, and dynamic supervision.

[0020] Optionally, the step of comparing the latest standard rule set with the pre-stored historical quality inspection rule base and generating a rule update instruction set based on the comparison results includes:

[0021] A unique identifier is generated for each rule in the latest standard rule set and the historical quality inspection rule base;

[0022] Based on the unique identifier, the latest standard rule set is matched with the corresponding rule in the historical quality inspection rule base;

[0023] Rules that fail to match are categorized into new rules, obsolete rules, or modified rules. Specifically, if a rule exists in the latest standard rule set but not in the historical quality inspection rule library, it is categorized as a new rule; if a rule exists in the historical quality inspection rule library but not in the latest standard rule set, it is categorized as an obsolete rule; and if the rule identifier matches but the content attributes are inconsistent, it is categorized as a modified rule.

[0024] The rule update instruction set is generated based on the classification results.

[0025] By adopting the above technical solution, a rule matching mechanism based on unique identifiers is constructed, and combined with set difference analysis methods, automatic identification and classification of quality inspection rule change types are achieved. This technical solution not only addresses the management challenges brought about by frequent rule updates but also improves system response speed while ensuring data consistency. The final generated rule update instruction set can be directly used to drive rule synchronization and deployment operations in downstream systems, thereby realizing intelligent and automated quality inspection rule management.

[0026] Optionally, based on the synchronously updated quality inspection rule base, the steps of performing automated quality audits on the target inefficient land use data to be processed and generating an original quality inspection result set include:

[0027] Obtain the synchronously updated quality inspection rule base and the target inefficient land use data to be processed;

[0028] Load the quality inspection rule base into the rule engine;

[0029] Iterate through each record of the target inefficient land use data, and match the constraint rules in the rule base for each field according to the field type; for basic constraint rules, call the preset general verification algorithm, and for complex logic rules, call the external script interface to perform verification.

[0030] Record field-level validation results, generate the original quality inspection result set, and aggregate and output it according to defect type.

[0031] By adopting the above technical solutions, the entire chain of automated processing from rule acquisition, engine loading, data traversal, rule matching to result output is realized. This not only effectively improves the efficiency and level of inefficient land use data quality management, but also provides reliable data support for the refined management and scientific decision-making of land resources.

[0032] Optionally, the step of parsing the defect distribution characteristics of the original quality inspection result set to generate a customizable quality inspection report includes:

[0033] Obtain the original quality inspection result set and the report template parameters configured by the user;

[0034] The defect distribution characteristics of the original quality inspection result set are analyzed to generate quantitative statistical indicators and problem classification data;

[0035] The quantitative statistical indicators are transformed into descriptive text paragraphs based on natural language generation technology;

[0036] Based on the report template parameters configured by the user, the descriptive text paragraphs and visualization chart components are combined to obtain a combination of text and graphics content;

[0037] The combined text and image content is packaged into a customizable quality inspection report according to a preset output format.

[0038] By adopting the above technical solutions, a complete automatic generation system for inefficient geological inspection reports was constructed. This solution not only significantly improves the efficiency and accuracy of report generation, but also enhances the readability and expressiveness of the reports through natural language generation and visualization technologies, thereby providing strong technical support for land resource management and data quality control.

[0039] Optionally, after the step of generating a customizable quality inspection report, the following may also be included:

[0040] Input the original quality inspection result set into the land resource management business system;

[0041] Receive related business indicator data returned by the land resource management business system; wherein, the related business indicator data includes at least one of land utilization rate, economic output volatility rate or planning conflict coefficient;

[0042] Based on the correlation weight between the associated business indicator data and the defect type, the business impact score of the defect record is calculated;

[0043] Defects whose business impact scores exceed a set scoring threshold will be marked as high-priority rectification items.

[0044] The customizable quality inspection report prominently displays the topological location map and associated land parcel information for high-priority rectification items.

[0045] By adopting the above technical solution, the original quality inspection result set is deeply integrated with the relevant business indicator data of the land resource management business system, constructing a defect scoring model based on the degree of business impact. This model enables the priority classification and visualization of rectification tasks. This solution effectively overcomes the problem of the disconnect between technical rules and business needs in traditional quality inspection processes. Through systematic data processing logic and an intelligent scoring mechanism, it achieves seamless integration from rule verification to business decision-making, providing strong technical support and data assurance for the governance of inefficient land use.

[0046] Optionally, the automated quality inspection method further includes:

[0047] Construct a target inefficient land use data table based on the target inefficient land use data to be processed;

[0048] Based on the target inefficient land use data table, data shards are dynamically divided in a distributed computing cluster;

[0049] The synchronously updated quality inspection rule base is divided into multiple rule subsets according to rule type and verification complexity;

[0050] Based on the field distribution characteristics of the data sharding, the rule subset is dynamically allocated to the corresponding computing nodes;

[0051] Field-level validation is performed in parallel across all compute nodes;

[0052] The defect datasets output by each computing node are aggregated, and conflict resolution is performed based on defect type and timestamp to generate a global quality inspection result set.

[0053] By adopting the above technical solution, the drawbacks of traditional serial processing methods, such as slow speed and difficulty in scaling, are overcome, and the side effects of blindly pursuing concurrency are cleverly avoided, thus achieving the goal of high-performance, high-quality, and high-reliability automated data governance.

[0054] Secondly, this application provides an automatic quality inspection system based on inefficient land use data, employing the following technical solution:

[0055] An automated quality inspection system based on inefficient land use data, the automated quality inspection system comprising:

[0056] The receiving module is used to receive user input of inefficient land use policy standard documents;

[0057] The semantic parsing module is used to perform semantic parsing on the inefficient land use policy standard document, identify rule entities and constraint logic, and generate a structured latest standard rule set;

[0058] The difference comparison module is used to compare the latest standard rule set with the pre-stored historical quality inspection rule library, and generate a rule update instruction set based on the difference comparison result.

[0059] The rule update module is used to dynamically update the quality inspection rule base according to the rule update instruction set and output the synchronously updated quality inspection rule base.

[0060] The quality inspection module is used to perform automated quality audits on the target inefficient land use data to be processed based on the synchronously updated quality inspection rule base, and generate an original quality inspection result set.

[0061] The quality inspection report generation module is used to analyze the defect distribution characteristics of the original quality inspection result set and generate a customizable quality inspection report.

[0062] Thirdly, this application provides a computer device, which adopts the following technical solution:

[0063] A computer device includes a memory, a processor, and a computer program stored in the memory, the processor executing the computer program to perform the steps of the method as described in the first aspect.

[0064] Fourthly, this application provides a computer-readable storage medium, which adopts the following technical solution:

[0065] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as in any of the methods in the first aspect.

[0066] In summary, this application includes at least one of the following beneficial technical effects: This application can automatically parse policy texts, extract structured rules, and implement high-quality data audits accordingly, thereby achieving an intelligent upgrade of the entire process from policy input to quality inspection output. Attached Figure Description

[0067] Figure 1 This is a first flowchart illustrating an automatic quality inspection method based on inefficient land use data, according to one embodiment of this application.

[0068] Figure 2 This is a second flowchart illustrating an automatic quality inspection method based on inefficient land use data, according to one embodiment of this application.

[0069] Figure 3This is a schematic diagram of the third process of an automatic quality inspection method based on inefficient land use data, according to one embodiment of this application.

[0070] Figure 4 This is a schematic diagram of the fourth process of an automatic quality inspection method based on inefficient land use data, according to one embodiment of this application.

[0071] Figure 5 This is a schematic diagram of the fifth process of an automatic quality inspection method based on inefficient land use data, according to one embodiment of this application.

[0072] Figure 6 This is a schematic diagram of the sixth process of an automatic quality inspection method based on inefficient land use data, which is one embodiment of this application.

[0073] Figure 7 This is a schematic diagram of the seventh process of an automatic quality inspection method based on inefficient land use data, according to one embodiment of this application. Detailed Implementation

[0074] To make the purpose, technical solution, and advantages of this application clearer, the following description is provided in conjunction with the appendix. Figure 1-7 The present application will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the application.

[0075] This application discloses an automatic quality inspection method based on inefficient land use data.

[0076] Reference Figure 1 An automated quality inspection method based on inefficient land use data, the quality inspection method includes:

[0077] Step S101: Receive the user-inputted standard document on inefficient land use policies;

[0078] Inefficient land use refers to the waste of land resources caused by reasons such as outdated industrial structure, low output efficiency, and unreasonable resource allocation during urban land use. Policy and standard documents are usually normative documents issued by local governments on how to define whether a plot of land belongs to inefficient land use, such as the "Management Measures for Redevelopment of Inefficient Industrial Land" or related implementation rules.

[0079] Understandably, these documents may exist in various formats such as PDF, Word, and TXT, often possessing high professionalism and structure, but lacking machine readability. Therefore, the core significance of this receiving step lies in establishing a unified information entry point for subsequent processing, enabling the centralized management of previously scattered and heterogeneous policy information, thereby laying the foundation for automated quality inspection.

[0080] Step S102: Semantic parsing of the inefficient land use policy standard document, identification of rule entities and constraint logic, and generation of a structured latest standard rule set;

[0081] Natural Language Processing (NLP) models can be used to perform deep language understanding and structure extraction tasks. Specifically, the NLP module can analyze policy text sentence by sentence to extract information such as field names (e.g., "annual output value," "plot ratio"), data table names (e.g., "enterprise basic information table," "land use attribute table"), and various logical constraints (e.g., "annual output value shall not be less than 5 million yuan / mu"). This semantic parsing is not a simple keyword matching process, but rather leverages pre-trained language models (e.g., BERT, RoBERTa) to achieve context-aware understanding, ensuring accurate capture of the implicit relationships between policy clauses.

[0082] The system then further organizes the extracted content into a series of structured rule tuples. Each tuple contains multiple core elements, such as table identifiers, field identifiers, numerical range constraints, and mandatory field identifiers. This structured representation not only facilitates rapid access and execution by computer programs but also provides a standardized basis for subsequent rule comparisons.

[0083] Step S103: Compare the latest standard rule set with the pre-stored historical quality inspection rule library, and generate a rule update instruction set based on the comparison results;

[0084] The key to this step lies in maintaining the consistency and timeliness of rule versions. To achieve this, the system calculates a unique hash value for each rule in the newly generated standard rule set and compares it one by one with existing rule items in the historical quality inspection rule base. If a rule's hash value cannot be matched in the old rule base, it indicates that it may be a newly added or modified rule; conversely, if some rules exist in the historical base but do not appear in the new set, they are marked as obsolete rules. This process is essentially an incremental update strategy, avoiding the performance loss caused by full replacement, while also ensuring the stability and security of the system. Based on the comparison results, the system will further generate a set of explicit operation instructions—the so-called "rule update instruction set"—to guide the actual operation in the next stage.

[0085] Step S104: Dynamically update the quality inspection rule base according to the rule update instruction set, and output the synchronously updated quality inspection rule base;

[0086] Dynamic updates emphasize the ability to respond to policy changes in real time, rather than traditional manual intervention or periodic batch imports. Upon receiving the aforementioned rule update command, the system will sequentially execute the corresponding insert, replace, or disable actions to ensure that the quality inspection rules in the current operating environment are always up-to-date. "Insert" corresponds to the entry of new rules into the database; "Replace" applies to cases where the same ID is retained but the content has changed; and "Disable" is a soft deletion measure for old rules that are no longer applicable, preventing accidental deletion from triggering a chain reaction. This approach effectively improves the flexibility and controllability of rule management while reducing operational costs.

[0087] Step S105: Based on the synchronously updated quality inspection rule base, perform automated quality audit on the target inefficient land use data to be processed, and generate the original quality inspection result set;

[0088] Specifically, the system already possesses a state-of-the-art, rigorously validated set of quality inspection rules, which has been deployed to the actual operating environment. The next task is to use these rules to conduct a comprehensive review of the data to be inspected. To this end, the system introduces a high-efficiency rule engine, which can be an open-source framework such as Drools or a self-developed lightweight interpreter. The rule engine is responsible for loading all effective rules and triggering the corresponding verification actions in sequence according to the set priority. During the verification process, on the one hand, basic constraints are checked using general algorithms, such as whether fields are empty, whether values ​​exceed the allowed range, and whether illegal characters exist; on the other hand, for more complex multi-field joint judgment scenarios (such as the dual judgment of "output value < 5 million and plot ratio < 0.6"), external script interfaces (such as Python functions or SQL stored procedures) are called for flexible extension. This ensures both the high efficiency and stability of basic functions and also takes into account the diversity of advanced business needs.

[0089] Step S106: Analyze the defect distribution characteristics of the original quality inspection result set and generate a customizable quality inspection report.

[0090] Although the preceding steps have completed the full data verification process, the original result set is often disordered, lengthy, and even difficult to understand intuitively, requiring further processing to generate valuable feedback. Therefore, at this stage, the system deeply mines and summarizes the original quality inspection results, extracting statistical indicators such as the percentage of error types, the distribution of problem areas, and the frequency of recurrence, thereby revealing the overall trend and development pattern of data quality issues. Furthermore, it integrates Natural Language Generation (NLG) technology to transform abstract defect records into easily readable text descriptions, such as "The land use of enterprise number X has a problem with missing annual output value." Simultaneously, users can choose different report template styles according to their preferences and insert chart elements such as bar charts and pie charts generated by the backend, making the report more expressive and persuasive. Finally, the system integrates all text and graphic materials, packages them into a formal electronic document according to the specified format, and exports and publishes it.

[0091] The above implementation constructs a closed-loop intelligent quality inspection system. It automatically parses inefficient land use policy standards using Natural Language Processing (NLP) and generates a structured rule set. Combined with a hash comparison mechanism, it dynamically updates the quality inspection rule base. A dual-mode rule engine (general algorithm + custom script) performs quality audits on standardized inefficient land use data. Finally, it outputs customizable graphic reports based on Natural Language Generation (NLG) technology. Compared to traditional methods relying on human intervention, this approach significantly improves the accuracy, timeliness, and sustainable development capabilities of quality inspection work, contributing to a comprehensive improvement in the level of refined natural resource management under the background of government digital transformation.

[0092] Reference Figure 2 As one implementation of step S102, the steps of semantically parsing the inefficient land use policy standard document, identifying rule entities and constraint logic, and generating a structured latest standard rule set include:

[0093] Step S201: Extract text content elements from the inefficient land use policy standard document to generate an initial semantic unit set;

[0094] Since policy documents related to inefficient land use are typically in formats such as PDF, Word, or TXT, which often contain non-plain text elements such as images, tables, headers, and footers, format stripping is necessary before semantic analysis. This process involves techniques such as Optical Character Recognition (OCR), Document Object Model (DOM) parsing, or text stream parsing to transform the original document into a unified plain text sequence. Subsequently, the text is tokenized to break it down into a series of lexical units with independent semantic functions, such as terms like "land parcel area," "floor area ratio," and "compliance with planning regulations."

[0095] On this basis, it is also necessary to introduce a stop word filtering mechanism to remove function words such as "de", "he", "shi" that have no actual semantic contribution, while retaining the key term vocabulary in the professional field to ensure that the subsequent entity recognition and relationship extraction processes can focus on the core content that truly carries the rule semantics. The final output result is a set composed of multiple preliminary semantic units, laying the foundation for the next step of entity annotation.

[0096] Step S202: Perform entity recognition on the initial semantic unit set, and annotate the field names, data table names, and constraint condition entities.

[0097] Among them, this step mainly relies on a named entity recognition model to complete. This model can be built based on deep learning architectures (such as BiLSTM-CRF, BERT-BiLSTM-CRF, etc.) and trained and optimized on the policy corpus in the land management field to adapt to the language expression habits in specific scenarios.

[0098] During this process, three core entity types need to be recognized: The first type is "data table name", which represents the logical structure name for storing relevant business data in the database, such as "Parcel Basic Information Table", "Construction Land Approval Record Table", etc.; the second type is "field name", which represents the basic attribute columns that make up the data table, such as "Parcel Number", "Planning Use Code", "Land Area", etc.; the third type is "constraint condition entity", which covers all keyword vocabularies used to limit the value range, type, or logical relationship of fields, such as "required", "greater than 0", "belonging to Class A land", "and", "or", etc. Through the recognition and annotation of these three types of entities, the semantic boundaries of the rule elements in the policy text can be initially established, providing a basic support for subsequent relationship modeling.

[0099] Step S203: Parse the logical relationships of the annotated entities based on a pre-trained language model, identify the dependency relationships between entities, and generate an enhanced semantic unit set.

[0100] Among them, the dependency relationships between entities include but are not limited to: the belonging relationship between the field name and the data table name, the conditional binding relationship between the field name and the value range, and the logical operator association between multiple fields.

[0101] In this embodiment, this step can use a pre-trained language model as the core computing engine, preferably the BERT architecture, and fine-tune the training based on it for text corpora in the land management field to give it stronger contextual understanding capabilities. By inputting the labeled entities into the model, the system can capture the syntactic dependency relationships and semantic dependency structures between different entities, thereby revealing the logical chains hidden in the text. Specifically, it mainly includes the following three typical dependency relationships: First, the "relationship between field name and data table name", that is, which specific data table structure a certain field belongs to, such as "land parcel number" should belong to "basic information table of land parcels"; second, the "conditional binding relationship between field name and value range", such as the "plot ratio" field must meet the numerical range restriction of "greater than or equal to 1.0 and less than or equal to 3.0"; and finally, the "logical operator association between multiple fields", such as "planned use code is Class A land" and "plot ratio is less than 2.0" are connected by "AND" to form a compound judgment condition.

[0102] Understandably, by identifying and modeling these dependencies, the original static semantic units are endowed with dynamic logical connection capabilities, thus forming a more expressive set of enhanced semantic units, providing richer semantic basis for subsequent rule-based structural transformation.

[0103] Step S204: Transform the enhanced semantic unit set into structured rule tuples to generate the latest standard rule set.

[0104] The aforementioned semantic enhancement results are further mapped into standardized rules that can be parsed and executed by computers. The core idea of ​​this process is to abstract each enhanced semantic unit into a triple structure (Subject-Predicate-Object), where Subject is the entity that actively applies the constraint (e.g., field name or table name), Predicate is the constraint type (e.g., required, range limitation, enumeration matching), and Object is the specific constraint value (e.g., "greater than 0", "residential land"), etc. This triple structure not only possesses good semantic clarity but also facilitates subsequent conversion to standard data formats such as JSON or XML, enabling integration into business systems for use by the rule engine.

[0105] Furthermore, to ensure the timeliness and traceability of the rule set, each rule entry should also include a version number and an effective timestamp field to facilitate system version management and rule update tracking. Through this series of structured transformation operations, the originally scattered and ambiguous policy texts are transformed into a highly standardized and computable rule system, significantly improving the enforceability of policy rules and the efficiency of system integration.

[0106] In the above implementation method, multi-level and multi-dimensional semantic parsing and logical modeling of inefficient land use policy standard documents are performed, realizing the automatic conversion from unstructured policy text to structured rule sets. This solution not only effectively solves the problems of low efficiency and error-proneness in traditional manual rule entry, but also enhances the ability to understand complex semantic relationships by introducing a pre-trained language model, making the entire parsing process more accurate and generalizable. Especially in the field of land resource management, this method helps to promote the digital implementation of policies and rules, providing solid data support and rule foundation for application scenarios such as intelligent approval, compliance review, and dynamic supervision.

[0107] Reference Figure 3 As one implementation of step S103, the step of comparing the latest standard rule set with the pre-stored historical quality inspection rule base and generating a rule update instruction set based on the comparison results includes:

[0108] Step S301: Generate a unique identifier for each rule based on the latest standard rule set and the historical quality inspection rule base;

[0109] The unique identifier can be generated either by calculating the hash value of the rule content itself, or by extracting and combining key rule attributes to form a string digest before encoding. The logic behind this design is that rule content is often highly complex and diverse; direct full-text comparison is not only computationally expensive but also difficult to guarantee the accuracy of consistency judgments. By abstracting the rule into a fixed-length or structured identifier, the matching complexity can be significantly reduced without affecting semantic integrity.

[0110] For example, for a quality inspection rule regarding land use classification, its fields include "use type," "maximum floor area ratio," and "building density limit." The system can extract these key attributes, concatenate them into a string, and then calculate the MD5 or SHA256 hash value to obtain the rule's unique identifier. The advantage of this method is that even if the rule description differs slightly (such as different line breaks or spaces), as long as the semantics remain unchanged, the identifier remains consistent, thus avoiding misjudgment.

[0111] Step S302: Match the latest standard rule set with the corresponding rules in the historical quality inspection rule base based on the unique identifier;

[0112] The core of this step lies in establishing a mapping relationship between the two rule sets, that is, identifying which rules remain consistent in the old and new versions, and which have changed. Since each rule has been assigned a unique identifier, the matching process is essentially transformed into intersection and difference operations between the two identifier sets. This process can be achieved at the database operation level using hash tables or indexing mechanisms for fast retrieval, greatly improving processing efficiency.

[0113] Furthermore, this matching mechanism implicitly includes a function to assess rule stability—if a rule's identifier remains unchanged across multiple versions, it indicates that its content is relatively mature and stable; conversely, if it changes frequently, its formulation logic may need to be re-examined. From an engineering implementation perspective, this mechanism also provides fundamental support for subsequent version control and rollback, making the rule management system more robust.

[0114] Step S303: Classify the rules that failed to match, and obtain the classification results as new rules, obsolete rules, or modified rules;

[0115] If a rule exists in the latest standard rule set but not in the historical quality inspection rule library, it is classified as a new rule; if a rule exists in the historical quality inspection rule library but not in the latest standard rule set, it is classified as an obsolete rule; if the rule identifier matches but the content attributes are inconsistent, it is classified as a changed rule.

[0116] Specifically, when a rule exists in the latest set of standard rules but not in the historical rule base, it indicates that the rule is a newly introduced quality control requirement and should be classified as a "new rule". Such rules usually reflect changes in policy adjustments, technological advancements, or business needs, such as the addition of quality inspection standards for land use in new industrial parks.

[0117] Conversely, if a rule exists only in the historical rule base and not in the latest standard rule set, it means that the rule has been phased out or replaced and should be classified as an "obsolete rule." The emergence of such rules often stems from the simplification and optimization of the standard system or the cleanup of redundant clauses.

[0118] Rules with the same identifier but different content attributes are classified as "modified rules," indicating that the original rules have been modified in terms of expression, constraints, or scope of application. For example, a rule that originally stipulated that the floor area ratio should not exceed 2.0 may be changed to not exceed 2.5 in the new version. Although the identifiers are the same, the business meaning has changed and must be treated as a modification. This classification mechanism not only helps to clearly define the rule evolution path but also provides a classification basis for the generation of subsequent update instructions.

[0119] Step S304: Generate a rule update instruction set based on the classification results.

[0120] In the above implementation, a rule matching mechanism based on unique identifiers is constructed, and combined with set difference analysis methods, automatic identification and classification of quality inspection rule change types are achieved. This technical solution not only addresses the management challenges brought about by frequent rule updates but also improves system response speed while ensuring data consistency. The final generated rule update instruction set can be directly used to drive rule synchronization and deployment operations in downstream systems, thereby realizing intelligent and automated quality inspection rule management.

[0121] Reference Figure 4 As one implementation of step S105, the step of performing automated quality audit on the target inefficient land use data to be processed, based on the synchronously updated quality inspection rule base, and generating the original quality inspection result set includes:

[0122] Step S401: Obtain the synchronously updated quality inspection rule base and the target inefficient land use data to be processed;

[0123] Synchronous updates imply a temporal consistency and version correspondence between the rule base and the data processing workflow. In practice, with adjustments to land management policies or updates to technical standards, the quality inspection rule base frequently changes. If there is a version mismatch between the rule base and the data to be processed, it may lead to inaccurate or invalid audit results. Therefore, establishing a mechanism to link rule versions with data batches, ensuring their synchronization on the logical timeline, is a prerequisite for guaranteeing audit effectiveness. For example, if a batch of land data was collected in the first quarter of 2024, the corresponding quality inspection rule base should also be the version that took effect at the same time. This avoids misjudgments caused by rules that are outdated or premature. This synchronization mechanism embodies the core concepts of version control and consistency management in data governance.

[0124] Step S402: Load the quality inspection rule library into the rule engine;

[0125] Among them, the rule engine, as a software component specifically designed to process business rules, works on the basis of a rule-data separation architecture. By abstracting business logic into a set of configurable rules, it decouples business processing logic from program code.

[0126] In this embodiment, the loading process of the rule engine is essentially the process of converting structured quality inspection rules into executable verification logic. This conversion process typically includes multiple sub-steps such as rule parsing, syntax validation, and dependency analysis. Internally, the rule engine often employs the Rete algorithm or a similar pattern matching mechanism to optimize rule execution efficiency. This algorithm, by constructing a rule condition network, avoids redundant computation and significantly improves the execution performance of large-scale rule sets. Furthermore, the rule engine also needs to have dynamic loading and hot update capabilities to adapt to the frequent changes in quality inspection rules. This design allows the system to complete rule updates without restarting the service, greatly improving system availability and maintenance efficiency.

[0127] Step S403: Iterate through each record of the target inefficient land use data and match the constraint rules in the rule base for each field according to the field type;

[0128] Among them, a preset general verification algorithm is called for basic constraint rules, and an external script interface is called to perform verification for complex logic rules;

[0129] Specifically, the core of this step lies in establishing a mapping relationship between field types and constraint rules. This mapping relationship is based on the concept of metadata management, that is, guiding the execution of validation logic through an abstract description of the data structure. Field type identification includes not only basic data types (such as strings, numbers, dates, etc.) but may also involve semantic types at the business level (such as administrative division codes, land use codes, etc.). The rule matching process is essentially a multi-dimensional filtering and screening process. The system needs to determine the applicable validation rules based on multiple dimensions such as the field's business attributes, data characteristics, and constraint strength. For example, for a numeric field representing land area, the system may need to simultaneously match multiple basic constraint rules such as non-emptyness checks, value range checks, and precision requirement checks. This refined rule matching mechanism ensures the comprehensiveness and relevance of the validation process, avoiding rule omissions or misuse.

[0130] Step S404: Record the field-level verification results, generate the original quality inspection result set, and aggregate and output it according to the defect type.

[0131] The field-level validation results records not only need to include basic status information such as whether the validation passed or failed, but also need to record in detail the key parameters in the validation process, such as validation rules, input data, expected results, and actual results. This information provides an important basis for subsequent problem analysis and quality traceability. The generation of the original quality inspection result set adopts a structured data organization method. By establishing a standardized data model for each defect record, the consistency and processability of the result data are ensured.

[0132] Furthermore, the process of aggregating and outputting defect types is essentially a secondary processing and value mining of quality data. Through statistical analysis of defect data, we can discover the distribution patterns of quality problems, identify high-frequency defect types, and assess the overall level of data quality. This aggregation analysis not only provides data support for quality improvement but also provides quantitative basis for management decisions. For example, by analyzing defect distribution histograms, we can identify data fields or business processes where quality problems are concentrated, thereby enabling targeted improvement measures to be developed.

[0133] The above implementation achieves fully automated processing from rule acquisition, engine loading, data traversal, rule matching to result output. This not only effectively improves the efficiency and level of inefficient land use data quality management, but also provides reliable data support for the refined management and scientific decision-making of land resources.

[0134] Reference Figure 5 As one implementation of step S106, the step of parsing the defect distribution characteristics of the original quality inspection result set and generating a customizable quality inspection report includes:

[0135] Step S501: Obtain the original quality inspection result set and the report template parameters configured by the user;

[0136] The original quality inspection result set typically originates from a systematic inspection of data related to inefficient land use. For example, it is a collection of defect records generated after verifying the completeness, consistency, and accuracy of land use status data, ownership data, and planning data. These records are usually in the form of structured data, such as JSON, CSV, or database table structures, containing information such as field names, defect types, defect descriptions, and the data table to which they belong.

[0137] Meanwhile, the user-configured report template parameters guide the structure and content layout of subsequent reports. For example, users may specify that the report should include visualization components such as defect distribution pie charts and defect rate heatmaps, as well as the organization of text paragraphs and the display order of charts. The key to this step is establishing a mapping relationship between data and templates, ensuring that subsequent processing stages can flexibly generate reports according to user intent.

[0138] Step S502: Analyze the defect distribution characteristics of the original quality inspection result set and generate quantitative statistical indicators and problem classification data;

[0139] This step involves performing multi-dimensional statistical analysis on the raw quality inspection data to extract key indicators reflecting the distribution patterns of defects. For example, by grouping and statistically analyzing data according to defect type fields, the frequency of occurrence for each type of defect (such as missing fields, incorrect field formats, logical conflicts, etc.) can be determined, thereby identifying the most common problem types in the system. Furthermore, by calculating the defect rate index (i.e., the proportion of defective records to the total number of records) for each data table, the quality level of different data tables can be quantified, helping users identify data sources with poor quality.

[0140] Furthermore, by analyzing the frequency of defect fields, the top N fields with the highest defect frequency and their associated rules can be identified, helping users focus on key issue fields and achieve accurate attribution of problems. Problem categorization data is typically aggregated by data table, grouping defect details by table name for easier structured presentation in subsequent reports. The logical basis of this step lies in transforming unstructured quality inspection results into interpretable and quantifiable statistical indicators through multi-dimensional aggregation and statistics of raw data, providing data support for subsequent natural language generation and visualization.

[0141] Step S503: Based on natural language generation technology, convert quantitative statistical indicators into descriptive text paragraphs;

[0142] This step involves more than simply filling numbers into a template. Instead, it constructs a predefined text template framework and combines it with text generation techniques from Natural Language Processing (NLP) to achieve a semantic expression of the statistical results. The text template framework contains multiple placeholders, each representing a key analytical dimension, such as "Defect type A accounts for X%" or "Field B is a high-frequency defect field."

[0143] In actual processing, the system fills the corresponding placeholders with the quantitative statistical indicators generated in the previous step, and optimizes the generated text at the syntactic and semantic levels through syntax tree reconstruction technology to ensure that the output text conforms to the norms of natural language expression. For example, when describing the distribution of defect types, the system will not only output "Field missing defects account for 35%", but may also generate more readable statements such as "Field missing defects account for the highest proportion of all defects, reaching 35%, indicating that data integrity issues are relatively prominent" through contextual semantic reasoning. The key to this step is to transform structured statistical results into easily understandable natural language, thereby improving the readability and usability of the report.

[0144] Step S504: Based on the report template parameters configured by the user, combine descriptive text paragraphs with visual chart components to obtain combined text and image content;

[0145] This step determines the insertion position and display order of text paragraphs and chart components in the report by parsing layout identifiers (such as section identifiers and component type identifiers) in the template parameters. The dynamic rendering of chart components relies on the visual mapping of statistical indicators; for example, mapping defect type distribution data to pie charts or bar charts, mapping defect rate data from data tables to heatmaps, or mapping trend data of high-frequency defect fields to line charts. These charts are typically generated based on SVG vector graphics or Canvas dynamic drawing technology to ensure compatibility and clarity across different output formats. The core of this step is to achieve semantic alignment and visual harmony between text and charts, ensuring that the report is both logically sound in its information expression and has a good visual presentation.

[0146] Step S505: Package the combined text and image content into a customizable quality inspection report according to a preset output format.

[0147] The system supports multiple output formats, such as PDF, HTML, and Word, each requiring specific encapsulation processing. For example, when generating PDF, the system needs to format the text and images, setting document attributes such as headers, footers, font styles, and paragraph spacing; when generating HTML, it needs to package the text, charts, and style resources to ensure correct rendering in the browser; and when generating Word, it needs to follow the structured specifications of .docx documents, organizing the content into paragraphs, tables, charts, and other objects. The technical challenge of this step lies in maintaining consistency and readability across different formats while meeting users' personalized needs for report formats.

[0148] In the above implementation, a complete automatic generation system for inefficient geological inspection reports was constructed. This solution not only significantly improves the efficiency and accuracy of report generation, but also enhances the readability and expressiveness of the reports through natural language generation and visualization technologies, thereby providing strong technical support for land resource management and data quality control.

[0149] Reference Figure 6 As a further implementation of the automated quality inspection method, after the step of generating a customizable quality inspection report, the method further includes:

[0150] Step S601: Input the original quality inspection result set into the land resource management business system;

[0151] The process of inputting the original quality inspection result set into the land resource management business system relies on the pre-set data interface protocol between the systems to ensure that the various defect records in the original quality inspection result set can be accurately matched with key fields such as land parcel identifiers and plot codes in the business system.

[0152] Step S602: Receive related business indicator data returned by the land resource management business system; wherein, the related business indicator data includes at least one of land utilization rate, economic output volatility, or planning conflict coefficient.

[0153] Specifically, the data on related business indicators is not obtained through manual surveys or external collection, but rather comes from business statistics related to the long-term operational status of each land parcel or plot in the land resource management business system. It includes, but is not limited to, key quantitative indicators such as land utilization rate, economic output volatility, and planning conflict coefficient.

[0154] Among them, land utilization rate reflects the degree of match between the actual use intensity of a specific plot of land per unit time and policy expectations; economic output volatility is used to characterize the stability of the plot of land in terms of economic output and the changing trend of its contribution to the regional economy; and planning conflict coefficient is used to measure the degree of deviation between the current use status of the plot of land and the overall planning and land use control requirements of its location. These indicator data are usually stored in time series form in the business system and are periodically updated and calculated through a preset business logic engine. Therefore, the return process is an automated call behavior between systems based on data identifier matching, ensuring the spatiotemporal consistency between indicator data and quality inspection defect records.

[0155] Step S603: Calculate the business impact score of the defect record based on the correlation weight between the associated business indicator data and the defect type;

[0156] In some embodiments, a correlation weight model is established between defect types and business metrics. By quantitatively analyzing the actual impact of various defects on key business metrics, a business-meaning score is assigned to each defect record. The correlation weights can be initially configured based on historical data analysis or expert experience rules, and dynamically optimized during system operation using machine learning algorithms.

[0157] For example, if a certain type of land use boundary deviation defect frequently leads to a significant decrease in land utilization rate in historical data, then this defect type will be assigned a higher correlation weight with the land utilization rate indicator. Similarly, if a certain type of land use mismatch defect often causes abnormal fluctuations in economic output, then its correlation weight with the economic output volatility indicator will also be increased accordingly. By substituting the defect type corresponding to each defect record and the associated business indicator data into this weighting model for weighted calculation, a business impact score for each defect record can be output. This scoring mechanism effectively solves the problem in traditional quality inspection processes that only focus on rule compliance while ignoring the degree of business impact, providing a scientific basis for subsequent rectification priority ranking.

[0158] Step S604: Mark defect records with business impact scores higher than the set score threshold as high-priority rectification items;

[0159] This process involves setting a scoring threshold to filter out defect records with business impact scores exceeding that threshold and mark them as high-priority rectification items. The scoring threshold can be dynamically adjusted based on system configuration parameters or business strategies. For example, in areas with scarce land resources, the threshold can be appropriately increased to focus on key issues, while in the early stages of policy implementation, the threshold can be lowered to comprehensively identify potential risks. This marking operation is not simply data filtering; rather, it achieves differentiated management of rectification tasks by adding a priority identifier field to the defect records. High-priority rectification items marked with this threshold will receive higher processing authority and resource allocation in subsequent business processes. For instance, they will be prioritized for handling by technical personnel in the task scheduling system and automatically pushed to senior management for decision-making in the approval process, thereby ensuring that key issues are responded to and effectively resolved in a timely manner.

[0160] Step S605: Highlight the topological location map and associated land parcel information of high-priority rectification items in the customizable quality inspection report.

[0161] Specifically, the system will highlight the topological location map of high-priority rectification items and their associated land parcel information in the report. The topological location map, generated by a GIS spatial data engine, can intuitively reflect the geographical distribution characteristics of each high-priority rectification item and their spatial relationships, facilitating managers to identify regional problems and systemic risks from a macro perspective. The associated land parcel information includes key attribute data such as parcel number, ownership unit, planned use, and historical rectification records, providing detailed data support for the implementation of specific rectification work.

[0162] Understandably, this visualization enhancement mechanism not only improves the readability and usability of quality inspection reports, but more importantly, it achieves a seamless connection between technical quality inspection and business rectification, enabling quality inspection results to be directly transformed into executable business instructions, significantly improving the intelligence level and decision-making efficiency of land resource management.

[0163] In the above implementation, the original quality inspection result set is deeply integrated with the relevant business indicator data of the land resource management business system to construct a defect scoring model based on the degree of business impact. This model enables the priority classification and visualization of rectification tasks. This solution effectively overcomes the problem of the disconnect between technical rules and business needs in traditional quality inspection processes. Through systematic data processing logic and an intelligent scoring mechanism, it achieves seamless integration from rule verification to business decision-making, providing strong technical support and data assurance for the governance of inefficient land use.

[0164] Reference Figure 7 As a further implementation of the automated quality inspection method, the automated quality inspection method also includes:

[0165] Step S701: Construct a target inefficient land use data table based on the target inefficient land use data to be processed;

[0166] This process involves organizing the original, collected, or imported data related to inefficient land use into structured database tables according to a standardized format. This data typically contains various types of information fields, such as structured attribute information like plot numbers, administrative division codes, land use status classifications, and names of ownership entities. It may also include spatial geographic information such as plot boundary coordinates and red-line vector graphics. The construction process involves metadata management, data cleaning, format conversion, and index optimization to ensure that the system can effectively read, parse, and process the data subsequently.

[0167] For example, in practical applications, a PostgreSQL database extended with PostGIS might be used to store geographic feature objects with geometric types, thereby supporting subsequent spatial queries and validations.

[0168] Step S702: Based on the target inefficient land use data table, dynamically divide the data into fragments in the distributed computing cluster;

[0169] Each data shard corresponds to at least one complete data table or a logically related subset of data tables. Dynamic partitioning means not simply dividing data blocks into fixed sizes, but making intelligent decisions after comprehensively considering factors such as the primary key distribution pattern, spatial relationships, and access frequency of the current data tables. By analyzing data heatmaps (i.e., map visualizations reflecting the data activity levels in different regions / time periods), hot and cold areas can be identified, thus determining whether they should be allocated to different physical nodes for execution, avoiding performance bottlenecks caused by excessive local load.

[0170] Furthermore, for a single ultra-large data table, spatial grid coding (such as GeoHash) or administrative boundary division can be introduced to perform fine-grained segmentation, so that each fragment is not only evenly distributed in number, but also maintains high cohesion and low coupling at the semantic level, which is conducive to improving parallel computing efficiency and cache hit rate.

[0171] Step S703: The synchronously updated quality inspection rule base is split into multiple rule subsets according to rule type and verification complexity;

[0172] Because different types and difficulty levels of rules require significantly different resource consumption during execution—for example, there is a clear difference between basic verification tasks that can be completed with only one traversal and advanced reasoning tasks that require multiple iterations or even external interface calls to reach a conclusion—it is necessary to classify and categorize them appropriately. For instance, a typical subset of rules might be dedicated to handling all the basic judgment logic regarding the validity of date fields, while others focus on more complex cross-table consistency comparisons.

[0173] Step S704: Based on the field distribution characteristics of data sharding, dynamically allocate rule subsets to the corresponding computing nodes;

[0174] Among them, the task assignment is based on feature vector matching, which involves constructing a rule vector representing the characteristics of the rule and a node vector describing the state of the node, and then using a mathematical model to find the optimal mapping relationship between the two.

[0175] Specifically, the Rule Vector includes four key metrics: the rule's category (e.g., syntax-based, business-based), the number of fields involved, expected computational overhead, and whether it depends on a specific spatial context. The Node Vector reflects the current available resources of each computing unit, including but not limited to parameters such as CPU utilization, remaining memory capacity, network communication latency, and disk I / O throughput. By minimizing the weighted Euclidean distance between the two, high-energy-consuming tasks can be routed to machines with high idle capacity as much as possible while maintaining overall load balancing, thereby achieving global optimization.

[0176] Step S705: Perform field-level validation in parallel on each computing node;

[0177] For ordinary structured fields, a mature rule engine framework is often used to scan the record content one by one to quickly locate violations of preset conditions. For special fields with geographical meaning, a dedicated topology verification algorithm is required for identification. Such programs are generally written by users and can accurately capture problems such as overlapping areal features and excessive gaps. As for the consistency review task involving fields related to multiple tables, it must rely on distributed transaction protocols to ensure that inconsistencies in state caused by concurrent modifications occur.

[0178] Step S706: Aggregate the defect datasets output by each computing node, resolve conflicts based on defect type and timestamp, and generate a global quality inspection result set.

[0179] Specifically, due to the highly parallel processing mode adopted in the early stage, it is inevitable that the same data will be repeatedly checked by multiple nodes at the same time. Therefore, it is necessary to establish a scientific and reasonable deduplication mechanism to eliminate redundant information interference.

[0180] For example, each verification activity is first uniquely identified, including the source node ID and the exact time it occurred. When two or more feedback reports point to the same entity but have different conclusions, their respective error types are compared. If the answer is yes, the earliest reported version is selected as the authoritative reference. Conversely, if conflicting viewpoints are found, a higher-level arbitration process is initiated, with a ruling made according to a pre-defined priority sequence. Furthermore, for cross-boundary issues requiring multi-party collaboration to determine ownership, the classic two-phase commit protocol is used to coordinate opinions and reach a consensus. The resulting global quality inspection report not only comprehensively covers all potential risk points but also possesses good consistency and credibility.

[0181] The above implementation overcomes the drawbacks of traditional serial processing methods, such as slow speed and difficulty in scaling, and cleverly avoids the side effects of blindly pursuing concurrency, thus achieving the goal of high-performance, high-quality, and high-reliability automated data governance.

[0182] This application also discloses an automatic quality inspection system based on inefficient land use data.

[0183] An automated quality inspection system based on inefficient land use data, comprising:

[0184] The receiving module is used to receive user input of inefficient land use policy standard documents;

[0185] The semantic parsing module is used to perform semantic parsing on the standard documents of inefficient land use policies, identify rule entities and constraint logic, and generate a set of the latest structured standard rules.

[0186] The difference comparison module is used to compare the latest standard rule set with the pre-stored historical quality inspection rule library, and generate a rule update instruction set based on the difference comparison results;

[0187] The rule update module is used to dynamically update the quality inspection rule base according to the rule update instruction set and output the synchronously updated quality inspection rule base.

[0188] The quality inspection module is used to perform automated quality audits on the target inefficient land use data to be processed, based on a synchronously updated quality inspection rule base, and generate an original quality inspection result set.

[0189] The quality inspection report generation module is used to analyze the defect distribution characteristics of the original quality inspection result set and generate a customizable quality inspection report.

[0190] The automatic quality inspection system based on inefficient land use data in this application embodiment can implement any of the above-mentioned automatic quality inspection methods, and the specific working process of each module in the automatic quality inspection system can refer to the corresponding process in the above-mentioned method embodiments.

[0191] In the several embodiments provided in this application, it should be understood that the provided methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for example, the division of a certain module is merely a logical functional division, and in actual implementation there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0192] This application also discloses a computer device.

[0193] Computer equipment, including memory, processor, and computer program stored in memory and executable on the processor, wherein the processor executes the computer program to implement the automatic quality inspection method based on inefficient land use data as described above.

[0194] This application also discloses a computer-readable storage medium.

[0195] A computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described above in any of the automated quality inspection methods based on inefficient land use data.

[0196] The computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device; the program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0197] It should be noted that the computer device and storage medium in the embodiments of this application are respectively electronic devices and storage media that apply the above-described automatic quality inspection method based on inefficient land use data. Therefore, all embodiments of the above-described automatic quality inspection method are applicable to the computer device and storage medium, and can achieve the same or similar beneficial effects. For the computer device / storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple; relevant details can be found in the descriptions of the method embodiments.

[0198] In this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0199] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, disclosure, and appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce a good effect.

[0200] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. An automatic quality inspection method based on low utility ground data, characterized in that, The quality inspection methods include: Receive user input of inefficient land use policy standard documents; Semantic parsing is performed on the aforementioned inefficient land use policy standard document to identify rule entities and constraint logic, and a structured latest standard rule set is generated; The latest standard rule set is compared with the pre-stored historical quality inspection rule base, and a rule update instruction set is generated based on the comparison results. The quality inspection rule base is dynamically updated according to the rule update instruction set, and the synchronously updated quality inspection rule base is output. Based on the synchronously updated quality inspection rule base, automated quality audits are performed on the target inefficient land use data to be processed, generating an original quality inspection result set; Analyze the defect distribution characteristics of the original quality inspection result set to generate a customizable quality inspection report; Based on the synchronously updated quality inspection rule base, the steps for performing automated quality audits on the target inefficient land use data to generate the original quality inspection result set include: Obtain the synchronously updated quality inspection rule base and the target inefficient land use data to be processed; Load the quality inspection rule base into the rule engine; Iterate through each record of the target inefficient land use data, and match the constraint rules in the rule base for each field according to the field type; for basic constraint rules, call the preset general verification algorithm, and for complex logic rules, call the external script interface to perform verification. Record field-level validation results, generate the original quality inspection result set, and aggregate and output it according to defect type.

2. The automatic quality inspection method based on inefficient land use data according to claim 1, characterized in that, The steps for semantically parsing the aforementioned inefficient land use policy standard document, identifying rule entities and constraint logic, and generating a structured, up-to-date standard rule set include: Extract text content elements from the inefficient land use policy standard document to generate an initial set of semantic units; Entity recognition is performed on the initial semantic unit set, and field names, data table names, and constraint entities are labeled. Based on a pre-trained language model, the logical relationships between labeled entities are parsed, the dependency relationships between entities are identified, and an enhanced semantic unit set is generated. The set of enhanced semantic units is transformed into structured rule tuples, generating the latest standard rule set.

3. The automatic quality inspection method based on inefficient land use data according to claim 1, characterized in that, The steps of comparing the latest standard rule set with the pre-stored historical quality inspection rule base and generating a rule update instruction set based on the comparison results include: A unique identifier is generated for each rule in the latest standard rule set and the historical quality inspection rule base; Based on the unique identifier, the latest standard rule set is matched with the corresponding rule in the historical quality inspection rule base; Rules that fail to match are categorized into new rules, obsolete rules, or modified rules. Specifically, if a rule exists in the latest standard rule set but not in the historical quality inspection rule library, it is categorized as a new rule; if a rule exists in the historical quality inspection rule library but not in the latest standard rule set, it is categorized as an obsolete rule; and if the rule identifier matches but the content attributes are inconsistent, it is categorized as a modified rule. The rule update instruction set is generated based on the classification results.

4. The automatic quality inspection method based on inefficient land use data according to claim 3, characterized in that, The steps for analyzing the defect distribution characteristics of the original quality inspection result set and generating a customizable quality inspection report include: Obtain the original quality inspection result set and the report template parameters configured by the user; The defect distribution characteristics of the original quality inspection result set are analyzed to generate quantitative statistical indicators and problem classification data; The quantitative statistical indicators are transformed into descriptive text paragraphs based on natural language generation technology; Based on the report template parameters configured by the user, the descriptive text paragraphs and visualization chart components are combined to obtain a combination of text and graphics content; The combined text and image content is packaged into a customizable quality inspection report according to a preset output format.

5. An automatic quality inspection method based on inefficient land use data according to any one of claims 1 to 4, characterized in that, Following the step of generating a customizable quality inspection report, the following steps are also included: Input the original quality inspection result set into the land resource management business system; Receive related business indicator data returned by the land resource management business system; wherein, the related business indicator data includes at least one of land utilization rate, economic output volatility rate or planning conflict coefficient; Based on the correlation weight between the associated business indicator data and the defect type, the business impact score of the defect record is calculated; Defects whose business impact scores exceed a set scoring threshold will be marked as high-priority rectification items. The customizable quality inspection report prominently displays the topological location map and associated land parcel information for high-priority rectification items.

6. The automatic quality inspection method based on inefficient land use data according to claim 1, characterized in that, The automated quality inspection method also includes: Construct a target inefficient land use data table based on the target inefficient land use data to be processed; Based on the target inefficient land use data table, data shards are dynamically divided in a distributed computing cluster; The synchronously updated quality inspection rule base is divided into multiple rule subsets according to rule type and verification complexity; Based on the field distribution characteristics of the data sharding, the rule subset is dynamically allocated to the corresponding computing nodes; Field-level validation is performed in parallel across all compute nodes; The defect datasets output by each computing node are aggregated, and conflict resolution is performed based on defect type and timestamp to generate a global quality inspection result set.

7. An automatic quality inspection system based on inefficient land use data, characterized in that, An automatic quality inspection system for performing an automatic quality inspection method based on inefficient land use data as described in any one of claims 1 to 6, the automatic quality inspection system comprising: The receiving module is used to receive user input of inefficient land use policy standard documents; The semantic parsing module is used to perform semantic parsing on the inefficient land use policy standard document, identify rule entities and constraint logic, and generate a structured latest standard rule set; The difference comparison module is used to compare the latest standard rule set with the pre-stored historical quality inspection rule library, and generate a rule update instruction set based on the difference comparison result. The rule update module is used to dynamically update the quality inspection rule base according to the rule update instruction set and output the synchronously updated quality inspection rule base. The quality inspection module is used to perform automated quality audits on the target inefficient land use data to be processed based on the synchronously updated quality inspection rule base, and generate an original quality inspection result set. The quality inspection report generation module is used to analyze the defect distribution characteristics of the original quality inspection result set and generate a customizable quality inspection report. The quality inspection module is configured as follows: Obtain the synchronously updated quality inspection rule base and the target inefficient land use data to be processed; Load the quality inspection rule base into the rule engine; Iterate through each record of the target inefficient land use data, and match the constraint rules in the rule base for each field according to the field type; for basic constraint rules, call the preset general verification algorithm, and for complex logic rules, call the external script interface to perform verification. Record field-level validation results, generate the original quality inspection result set, and aggregate and output it according to defect type.

8. A computer device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as claimed in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer program is stored that can be loaded by a processor and executed as described in any one of claims 1 to 6.

Images