Soil testing and detection data intelligent management system based on big data

Abstract

The embodiment of the application provides a soil inspection and detection data intelligent management system based on big data. The system comprises a data acquisition unit, a data preprocessing unit, a model analysis unit and an intelligent management and control unit. The data acquisition unit is configured to acquire multi-dimensional soil inspection and detection data of a region to be managed. The data preprocessing unit is electrically connected with the data acquisition unit and is configured to preprocess data in an original database. The model analysis unit is electrically connected with the data preprocessing unit and is configured to judge whether real-time collected soil inspection and detection data is abnormal data based on an output result of a management model. The intelligent management and control unit is electrically connected with the model analysis unit. The system can realize integrated, standardized and intelligent management of soil detection data from acquisition, processing and analysis to management, early warning and storage, improve soil detection data management efficiency and quality, and provide accurate support for soil environmental quality evaluation, pollution prevention and control and decision-making.

Description

Technical Field

[0001] This application relates to the field of data management technology, and in particular to an intelligent management system for soil testing and inspection data based on big data. Background Technology

[0002] Soil environmental monitoring is a core component of ecological and environmental protection. Soil testing and inspection data serves as a crucial basis for soil quality assessment, pollution source tracing, and pollution prevention and control decisions. The standardization, precision, and efficiency of soil testing and inspection data management directly determine the effectiveness of soil environmental supervision and remediation. With the comprehensive advancement of soil environmental protection in my country, the coverage of soil testing has continuously expanded, extending from industrial parks and mining areas to various regions such as agricultural land and construction land. Testing items have also expanded from basic physicochemical indicators to multiple pollutant indicators such as heavy metals, volatile organic compounds (VOCs), and semi-volatile organic compounds (SOCs). Soil testing and inspection data are gradually exhibiting characteristics of multi-dimensionality, large scale, strong temporal and spatial characteristics, and prominent professional attributes, significantly increasing the requirements for the professionalism and intelligence of data management systems.

[0003] Currently, the management of soil testing data still relies primarily on traditional methods, which depend on manual recording, decentralized storage, and simple statistical analysis. This approach has not yet been deeply integrated with big data and machine learning technologies, and is therefore unable to meet the current needs for large-scale, multi-dimensional soil testing data management. In practical applications, this has revealed many problems that urgently need to be addressed, specifically in the following aspects: First, data collection and storage are fragmented, forming "data silos." Under the traditional management model, multi-dimensional information such as soil physicochemical indicators, pollutant concentrations, sampling geospatial information, testing equipment operation data, and historical testing data are mostly recorded in different spreadsheets, paper documents, or simple management systems. There is a lack of a unified collection architecture and a standardized storage system. Effective relationships are not established between the data. When retrieving, integrating, and analyzing data, manual cross-file and cross-system sorting is required, which is not only time-consuming and labor-intensive, but also prone to data mismatch problems, significantly reducing the efficiency of data utilization.

[0004] Secondly, the data preprocessing methods are crude, making it difficult to guarantee the authenticity and validity of the data. Preprocessing of soil testing data is the foundation for subsequent analysis, but in traditional management, data preprocessing is mostly done manually, lacking a professional automated processing mechanism: invalid data resulting from sampling operation errors, testing equipment malfunctions, and manual recording errors are not thoroughly removed; to address data missing issues, a simple global mean method is often used for completion, without considering the professional characteristics of soil testing to differentiate between professional data such as physicochemical indicators and pollutant indicators and basic information such as geographical and temporal information; at the same time, testing data of different dimensions are not uniformly normalized, resulting in data disorder and low standardization, creating serious hidden dangers for subsequent data analysis and quality assessment.

[0005] Third, weak data analysis capabilities prevent in-depth data mining and accurate anomaly identification. Traditional management models can only perform simple statistics and queries, failing to build targeted analysis models based on soil testing industry standards and professional rules. The application of big data technology and machine learning algorithms is severely inadequate, making it impossible to uncover the inherent correlations between testing indicators, the spatiotemporal trends of testing data, or accurately identify data anomalies. Anomaly investigation relies heavily on manual comparison and analysis, which is not only inefficient but also prone to misjudgments and omissions due to subjective human judgment. Abnormal data cannot be detected and addressed in a timely manner, thus interfering with the accuracy of soil quality assessment results.

[0006] Fourth, soil quality assessment is one-sided, and data management is disconnected from environmental supervision. Traditional soil quality grading often relies solely on the compliance rate of a single indicator, failing to consider key factors such as data completeness and indicator trends. As a result, the assessment results cannot fully reflect the actual state of the soil environment. Furthermore, a systematic tiered early warning mechanism has not been established, early warning information is poorly disseminated, and data sharing with soil environmental supervision platforms at all levels is impossible. This prevents regulatory departments from promptly grasping the soil pollution situation and carrying out targeted pollution prevention and control work. In addition, traditional data management only focuses on data storage, failing to provide professional decision-making support for optimizing soil testing sites and developing contaminated soil remediation plans, thus failing to meet the actual needs of precise soil environmental prevention and scientific governance.

[0007] In summary, traditional soil testing and inspection data management models are no longer adequate to meet the evolving needs of current soil environmental monitoring. Problems such as data fragmentation, inefficient processing, superficial analysis, one-sided assessments, disconnected supervision, and difficulties in tracing the source of pollution significantly reduce the value of soil testing data, failing to provide effective data support for soil environmental quality assessment, pollution prevention and control, and scientific governance. Therefore, developing a system that combines big data technology with the professional needs of the soil testing industry to achieve standardized and intelligent management of soil testing and inspection data throughout the entire process from collection, preprocessing, analysis to control, early warning, and storage has become a pressing technical challenge in the field of soil environmental monitoring. Summary of the Invention

[0008] Based on this, it is necessary to provide a big data-based intelligent management system for soil testing and inspection data to address the above-mentioned technical issues, including a data acquisition unit, a data preprocessing unit, a model analysis unit, and an intelligent control unit. The data acquisition unit is configured to collect multi-dimensional soil testing data from the area to be managed and construct a raw soil testing database. The data preprocessing unit is electrically connected to the data acquisition unit and is configured to preprocess the data in the original database, build an intelligent management model for soil testing data based on big data technology, and input the preprocessed data into the management model. The model analysis unit is electrically connected to the data preprocessing unit and is configured to determine whether the real-time collected soil testing data is abnormal based on the output of the management model. The intelligent control unit is electrically connected to the model analysis unit and is configured to perform corresponding correction and completion operations on the soil testing data according to the anomaly type when the data is determined to be abnormal. When the data is determined to be non-abnormal, the soil testing data is intelligently classified, archived, and trend analyzed. The soil quality grade of the area to be managed is determined in conjunction with the soil testing industry standards, and a data management report is generated.

[0009] In one embodiment, the data acquisition unit is configured to acquire physicochemical index data, pollutant detection data, sampling geospatial information data, detection equipment operation data, and historical inspection and testing data of the soil in the area to be managed. It is also configured to classify the multi-dimensional data into structured and unstructured data, build a raw soil testing database based on a distributed storage architecture, and add basic tags such as sampling time, sampling location, and testing items to each data entry.

[0010] In one embodiment, the data preprocessing unit is configured to perform data cleaning operations on the raw data to remove invalid data caused by sampling operation errors, detection equipment failures, or human recording errors. It is also configured to complete the missing values ​​of the cleaned data, using the mean interpolation method of soil of the same type in the same region for soil testing professional indicators, and the source verification and completion method for geographic information and time information; Simultaneously, it is configured to perform dimensional unification and normalization on the completed data to obtain standardized soil testing data. The preprocessing formula is: in, For standardized data, This is the original test data. It is the minimum value of the test data for the same project. This represents the maximum value of the test data for the same project.

[0011] In one embodiment, the model analysis unit is configured to integrate machine learning algorithms with soil testing and inspection industry rules to construct a multi-dimensional data association intelligent management model. The model input is standardized soil testing data, and the model output is data association, indicator change trends, data quality rating, and anomaly identification results. It is also configured to use historical soil testing data of the area to be managed to initially train the management model, and combine it with real-time collected soil testing data to perform incremental training, dynamically update the model parameters, and improve the model's recognition and analysis accuracy. It is also configured to build a dedicated knowledge graph for soil testing, which will perform multi-dimensional correlation modeling of sampling geospatial information, soil type, testing indicators, pollutant type, pollution level, soil improvement measures and historical testing data to form a visual knowledge association network. The knowledge graph will be dynamically updated in conjunction with real-time collected soil testing data to maintain the timeliness of the graph information.

[0012] In one embodiment, the model analysis unit is configured to calculate the correlation value between real-time detection data and historical data in the same area, as well as the analysis residual value of real-time detection data, through a management model. It is also configured to preset correlation thresholds and residual thresholds for soil testing data, compare the calculated values ​​of real-time data with the thresholds to determine whether the data is abnormal, and identify the type of abnormality as sampling abnormality, detection abnormality, data transmission abnormality, or indicator mutation abnormality.

[0013] In one embodiment, the intelligent control unit is configured to, when identifying an anomaly type as sampling anomaly or detection anomaly, trigger a supplementary sampling instruction for the corresponding sampling point or a re-inspection instruction for the detection item, replace the abnormal data with the valid data after supplementary sampling and re-inspection, and update the original database; configured to, when identifying an anomaly type as data transmission anomaly, perform data retransmission and verification operations until the data transmission is error-free; and configured to, when identifying an anomaly type as indicator mutation anomaly, mark the data and associate it with data from surrounding points in the same area for verification to confirm whether it is a change in actual soil indicators.

[0014] In one embodiment, the intelligent control unit is configured to classify and archive non-abnormal data in multiple levels according to soil use type, testing items, and pollutant type, establish a data association retrieval system, and support accurate retrieval based on multiple conditions such as sampling point, time, and indicators; it is also configured to perform trend analysis on soil testing indicators at each sampling point based on big data time series analysis technology, obtain the rate of change, cycle of change, and correlation of indicators, and generate a visualized indicator change trend chart. It is also configured to retrieve a soil testing-specific knowledge graph constructed by the model analysis unit, enabling intelligent knowledge retrieval of the graph and intelligent reasoning based on real-time testing data and graph association rules, providing knowledge-level decision support for optimizing soil testing points in the area to be managed, formulating targeted soil improvement plans, and predicting pollution trends.

[0015] In one embodiment, the intelligent control unit is configured to calculate a comprehensive soil quality score based on standardized data of various soil testing indicators, combined with industry standard thresholds. The scoring formula is as follows: in, For the comprehensive soil quality score, To achieve the target compliance rate, For data integrity, To assess the health of the indicator trend, These are the weighting coefficients, and All values ​​are not zero; it is also configured to generate a data management report based on the comprehensive score, which includes a quality assessment of soil testing data in the area to be managed, analysis of index changes, data archiving status, and recommendations for subsequent testing.

[0016] In one embodiment, the intelligent control unit is configured to pre-configure a first preset score, a second preset score, and a third preset score for soil quality, wherein the first preset score < the second preset score < the third preset score; The system is configured to determine the soil quality grade as excellent when the comprehensive soil quality score is greater than the third preset score, good when the comprehensive score is greater than the second preset score and less than or equal to the third preset score, medium when the comprehensive score is greater than the first preset score and less than or equal to the second preset score, and poor when the comprehensive score is less than or equal to the first preset score. It is also configured to generate targeted soil improvement suggestions based on the type of pollutants and the trend of indicator changes, combined with the reasoning results of knowledge graph, when the soil quality level is medium or poor, and trigger graded early warning prompts.

[0017] In one embodiment, the system further includes an early warning push unit and a data storage unit. The early warning push unit is electrically connected to the intelligent control unit and is configured to receive graded early warning prompts when the soil quality level is medium or poor, push soil improvement suggestions and detection adjustment strategies, and synchronize the early warning information to the soil environmental monitoring platform. The data storage unit is electrically connected to the data acquisition unit and the intelligent control unit. It is configured to adopt a cloud storage architecture to store raw test data, standardized data, data management reports, soil quality grade results, early warning information and soil test-specific knowledge graph data throughout their entire lifecycle. It supports data traceability, retrieval and backup. The data storage unit is also equipped with an access permission management module to enable hierarchical data access for staff at different levels.

[0018] The beneficial effects of this invention are: 1. Standardized end-to-end processing of multi-dimensional soil testing data has been achieved, effectively improving the authenticity and validity of the data. Through the data preprocessing unit, targeted cleaning, missing value completion, and normalization are performed on the raw data. Dedicated completion strategies are designed for professional soil testing indicators to accurately remove invalid data generated by sampling, equipment, and human errors. Compared with traditional manual processing methods, this not only achieves full automation of data preprocessing but also significantly improves the efficiency of soil testing data. It completely solves the problems of data disorder, inconsistent dimensions, and arbitrary handling of missing values ​​in traditional management, laying a high-quality data foundation for subsequent data analysis.

[0019] 2. An intelligent analysis model tailored to the needs of the soil testing industry has been constructed, enabling accurate identification and targeted handling of data anomalies. The system's model analysis unit integrates machine learning algorithms with soil testing industry standards and professional rules to build an intelligent management model. Coupled with a dynamic training mechanism, the model continuously adapts to real-time testing data. At the same time, a two-dimensional quantitative judgment system is designed to accurately identify abnormal data and can also perform refined source tracing of anomaly types. Dedicated correction and completion strategies are configured for different anomaly types. Compared with the traditional manual anomaly screening method, this not only improves the anomaly identification accuracy to over 90%, but also significantly shortens the anomaly handling time, avoiding interference from abnormal data to soil quality assessment from the source and ensuring the reliability of data analysis results.

[0020] 3. A dedicated knowledge graph for soil testing has been built, achieving an upgrade from "data management" to "knowledge empowerment," providing scientific decision-making support for soil environmental governance. This system breaks through the limitations of traditional data management, which only focuses on data storage and simple statistics. Through the knowledge graph, it links and models multi-dimensional information such as sampling points, soil types, pollutants, and remediation measures, enabling knowledge-based retrieval and intelligent reasoning. Based on real-time testing data, it can provide professional knowledge support for optimizing soil testing points, formulating targeted remediation plans, and predicting pollution trends, allowing big data analysis results to be truly applied to actual decision-making in soil environmental governance and enhancing the utilization value of data.

[0021] 4. The system achieves refined classification, archiving, and efficient retrieval of soil testing data, significantly improving data management and utilization efficiency. The intelligent control unit of this system classifies and archives non-abnormal data at multiple levels according to soil usage type, testing items, and pollutant types. It has established a multi-condition accurate retrieval system based on sampling location, time, and indicators. Compared with traditional file-based and tabular data storage methods, the retrieval response time is greatly shortened, solving the problems of difficult data retrieval and low reusability in traditional management. At the same time, time series analysis technology is used to uncover the indicator change patterns behind the data, allowing the value of the testing data to be fully explored.

[0022] 5. A multi-dimensional comprehensive soil quality evaluation system has been established, making soil quality grading more scientific and comprehensive. This system abandons the traditional single model of judging soil quality solely based on the compliance rate of indicators. It calculates a comprehensive soil quality score based on three dimensions: compliance rate of indicators, data completeness, and indicator trend health. It divides the soil quality into refined grades, so that soil quality assessment is no longer limited to static indicator testing, but also takes into account the completeness of data and the dynamic trend of indicator changes. The assessment results are more in line with the actual conditions of the soil environment, providing a precise basis for the graded management and control of soil environmental quality.

[0023] 6. A mechanism linking tiered early warning and regulatory platforms has been established, improving the timeliness and coordination of soil pollution prevention and control. This system sets tiered early warnings for different soil quality levels, pushes early warning information through multiple channels, and simultaneously synchronizes the early warning information to soil environmental regulatory platforms at all levels. This achieves seamless linkage between monitoring data management and environmental supervision. Compared with the problems of delayed early warnings and poor information transmission in traditional management, it significantly shortens the early warning response time, enabling regulatory departments to grasp the soil pollution status in a timely manner, take targeted prevention and control measures, and effectively curb the spread of soil pollution. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the system composition of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0026] like Figure 1 As shown, the intelligent management system for soil testing and inspection data based on big data includes a data acquisition unit, a data preprocessing unit, a model analysis unit, and an intelligent control unit. The data acquisition unit is configured to collect multi-dimensional soil testing data from the area to be managed, constructing a raw soil testing database. The data preprocessing unit, electrically connected to the data acquisition unit, is configured to preprocess the data in the raw database, build an intelligent management model for soil testing data based on big data technology, and input the preprocessed data into the management model. The model analysis unit, electrically connected to the data preprocessing unit, is configured to determine whether the real-time collected soil testing data is abnormal based on the output of the management model. The intelligent control unit, electrically connected to the model analysis unit, is configured to perform corresponding correction and completion operations on the soil testing data according to the type of abnormality when the data is determined to be abnormal, and to perform intelligent classification, archiving, and trend analysis on the soil testing data when the data is determined to be non-abnormal, and to determine the soil quality level of the area to be managed in conjunction with soil testing industry standards, while generating a data management report.

[0027] Specifically, the system is divided into four main units based on the core data management process. Each unit focuses on its specific task, with no functional overlap, precisely matching industry-specific needs. This addresses the fragmentation problem inherent in traditional data management and avoids the issues of inefficient and inaccurate processing caused by a single module handling multiple tasks. The units are electrically connected sequentially, enabling efficient bidirectional flow of data and instructions, forming a closed-loop management system from data collection to application. This eliminates the need for manual data transfer, improving management efficiency and avoiding data mismatches and omissions caused by manual transmission. The data preprocessing unit serves as a pre-processing step for the model analysis unit, directly inputting standardized data into the management model to avoid interference from messy and distorted raw data. The model analysis unit focuses on the core task of anomaly detection, addressing the subjectivity and high rates of false positives and false negatives inherent in traditional manual judgment. The intelligent control unit develops specific handling strategies for both abnormal and non-abnormal data. Abnormal data is corrected and supplemented to ensure its authenticity and validity, while non-abnormal data undergoes classification, archiving, trend analysis, quality rating, and report generation, resolving the problem of traditional data management where data is simply stored but not used, resulting in low utilization value. The four core units are independent functional modules without strong coupling limitations. The operating parameters of each unit can be flexibly adjusted to adapt to different regional and type-specific testing needs. New functional units can also be added directly to expand functionality without reconstructing the core architecture, significantly improving the system's adaptability and future scalability. From data acquisition, preprocessing, anomaly detection to data handling and quality assessment, the entire process is automatically completed through the coordinated operation of each unit. Only a small number of personnel are needed for parameter settings and system maintenance, significantly reducing labor costs and completely avoiding various errors caused by traditional manual operations, achieving a dual improvement in management efficiency and accuracy. The anomaly detection results of the model analysis unit directly support the quality assessment work of the intelligent control unit, making soil quality level determination and data management report generation core control functions. This ensures that data analysis always revolves around actual business objectives, solving the problem of the disconnect between traditional data analysis and practical application, and truly implementing big data technology to serve the actual needs of soil testing data management.

[0028] The data acquisition unit is configured to acquire physicochemical index data, pollutant detection data, sampling geospatial information data, detection equipment operation data, and historical inspection and testing data of the soil in the area to be managed. It is also configured to classify multi-dimensional data into structured and unstructured data, build a raw soil testing database based on a distributed storage architecture, and add basic tags such as sampling time, sampling point, and testing items to each data entry.

[0029] Specifically, multi-dimensional, full-volume data collection comprehensively gathers soil testing information, laying a solid and detailed data foundation for subsequent analysis and judgment. Data is stored categorized by structure type to adapt to the management needs of different data characteristics, making raw data storage more organized and orderly. A distributed storage architecture is adopted to improve the database's storage capacity, stability, and scalability, adapting to the management needs of large-scale soil testing data. Dedicated basic tags are added to the data to achieve accurate identification and association, providing convenient conditions for subsequent rapid data retrieval, classification analysis, and end-to-end traceability. Standardized initial management of data is achieved from the collection end, reducing redundant operations in subsequent data processing and improving overall data management efficiency.

[0030] The data preprocessing unit is configured to perform data cleaning on the raw data, removing invalid data caused by sampling errors, testing equipment malfunctions, and manual recording errors. It is also configured to complete the cleaned data by imputing missing values, using the mean interpolation method for soil testing indicators of the same type and region, and the source verification and imputation method for geographic and temporal information. Simultaneously, it is configured to perform dimensional unification and normalization on the imputed data to obtain standardized soil testing data. The preprocessing formula is: in, For standardized data, This is the original test data. It is the minimum value of the test data for the same project. This represents the maximum value of the test data for the same project.

[0031] Specifically, precise cleaning of invalid data eliminates interference from sampling, equipment, and human error at the source, ensuring data authenticity and validity. Missing values ​​are filled in differently according to their type, aligning with the professional attributes of soil testing and avoiding the distortion caused by a single filling method, thus improving data completeness. Dimensional unification and normalization eliminate dimensional differences between different testing indicators, making the data comparable and analyzable. Standardized processing outputs well-organized data, providing a high-quality data foundation for subsequent model analysis and anomaly detection, ensuring the accuracy of analytical results. The entire preprocessing process is automated, replacing manual operation, reducing human error, and significantly improving data processing efficiency.

[0032] The model analysis unit is configured to integrate machine learning algorithms with soil testing and inspection industry rules to construct a multi-dimensional data-linked intelligent management model. The model input is standardized soil testing data, and the model output includes data correlation, indicator change trends, data quality rating, and anomaly identification results. It is also configured to use historical soil testing data from the area to be managed for initial training of the management model, and to perform incremental training using real-time collected soil testing data to dynamically update model parameters and improve the model's recognition and analysis accuracy. Furthermore, it is configured to build a soil testing-specific knowledge graph, which performs multi-dimensional correlation modeling of sampled geospatial information, soil type, testing indicators, pollutant type, pollution level, soil remediation measures, and historical testing data to form a visualized knowledge network. This knowledge graph is dynamically updated using real-time collected soil testing data to maintain the timeliness of the graph information.

[0033] The model analysis unit is configured to calculate the correlation value between real-time detection data and historical data in the same area, as well as the analysis residual value of real-time detection data, through the management model; it is also configured to preset the correlation threshold and residual threshold of soil detection data, compare the calculated value of real-time data with the threshold, determine whether the data is abnormal, and identify the type of abnormality as sampling abnormality, detection abnormality, data transmission abnormality, or indicator mutation abnormality.

[0034] Specifically, the model integrates algorithms and industry rules to align with the professional attributes of soil testing. The model outputs multi-dimensional analysis results, providing comprehensive data support for subsequent management and control. A dual-mode training approach (initial + incremental) is employed, dynamically updating parameters to continuously adapt the model to real-time testing data and improve analysis and identification accuracy. A dedicated knowledge graph for soil testing is built and dynamically updated, enabling multi-dimensional information association modeling and visualization, ensuring the timeliness of graph information, and uncovering the knowledge value behind the data. Anomalies are identified using a combination of correlation degree and residual value indicators with thresholds, ensuring quantitative and objective judgment standards and significantly improving the accuracy of anomaly identification. Four specific types of data anomalies are accurately identified, providing a clear basis for targeted handling of anomalies and avoiding blind anomaly management.

[0035] The intelligent control unit is configured to, when identifying an anomaly type as sampling anomaly or detection anomaly, trigger a supplementary sampling instruction for the corresponding sampling point or a re-inspection instruction for the detection item, replace the abnormal data with the valid data after supplementary sampling and re-inspection, and update the original database; it is configured to, when identifying an anomaly type as data transmission anomaly, perform data retransmission and verification operations until the data transmission is error-free; and it is configured to, when identifying an anomaly type as indicator mutation anomaly, mark the data and correlate it with data from surrounding points in the same area for verification to confirm whether it is a change in actual soil indicators.

[0036] The intelligent management and control unit is configured to classify and archive non-abnormal data at multiple levels according to soil use type, testing items, and pollutant types, and establish a data association retrieval system that supports precise retrieval based on multiple conditions such as sampling point, time, and indicators. It is also configured to perform trend analysis on soil testing indicators at each sampling point based on big data time-series analysis technology, obtaining the rate of change, cycle of change, and correlation patterns of indicators, and generating a visualized indicator trend chart. Furthermore, it is configured to retrieve a soil testing-specific knowledge graph constructed by the model analysis unit, enabling intelligent knowledge-based retrieval of the graph and intelligent reasoning based on real-time testing data and graph association rules, providing knowledge-level decision support for optimizing soil testing points in the managed area, formulating targeted soil improvement plans, and predicting pollution trends.

[0037] The intelligent control unit is configured to calculate a comprehensive soil quality score based on standardized data from various soil testing indicators, combined with industry standard thresholds. The scoring formula is as follows: .in, For the comprehensive soil quality score, To achieve the target compliance rate, For data integrity, To assess the health of the indicator trend, These are the weighting coefficients, and All values ​​are not zero; it is also configured to generate a data management report based on the comprehensive score, which includes a quality assessment of soil testing data in the area to be managed, analysis of index changes, data archiving status, and recommendations for subsequent testing.

[0038] The intelligent control unit is configured to pre-configure a first preset score, a second preset score, and a third preset score for soil quality, wherein the first preset score < the second preset score < the third preset score; it is configured to determine the soil quality level as excellent when the comprehensive soil quality score is greater than the third preset score, as good when the comprehensive score is greater than the second preset score and less than or equal to the third preset score, as medium when the comprehensive score is greater than the first preset score and less than or equal to the second preset score, and as poor when the comprehensive score is less than or equal to the first preset score; it is also configured to generate targeted soil improvement suggestions based on the type of pollutants and the trend of indicator changes, combined with the reasoning results of the knowledge graph, and trigger a graded early warning prompt when the soil quality level is medium or poor.

[0039] Specifically, abnormal data is handled precisely according to type, with different anomalies matched with dedicated operations such as re-sampling and re-testing, retransmission verification, and surrounding area review. The database is updated after correction, ensuring the authenticity and validity of soil testing data from the source. Non-abnormal data is classified and archived at multiple levels, and a multi-condition retrieval system is established to improve data management and retrieval efficiency. Time-series analysis uncovers the patterns of indicator changes and generates visual trend charts. Combined with knowledge graph intelligent reasoning, this provides professional knowledge support for optimizing testing sites, formulating improvement plans, and predicting pollution trends, fully leveraging the value of the data. A comprehensive soil quality score is calculated by integrating indicator compliance rate, data integrity, and trend health. The rating is quantified using industry standards and weighting coefficients, ensuring a scientific and comprehensive basis for judgment and avoiding the one-sidedness of single-indicator ratings. Four levels of soil quality are determined based on the comprehensive score. Medium / poor levels automatically trigger tiered early warnings, and targeted improvement suggestions are generated based on knowledge graph reasoning results, making soil pollution prevention and control more precise and achieving coordinated implementation of management and early warning. It automatically generates management reports that integrate data quality assessment, indicator analysis, archiving status, and testing recommendations, making data management results visible and actionable, and providing direct and comprehensive decision-making basis for soil testing and environmental remediation.

[0040] The system also includes an early warning push unit and a data storage unit. The early warning push unit is electrically connected to the intelligent control unit and is configured to receive graded early warning prompts when the soil quality level is medium or poor, push soil improvement suggestions and testing adjustment strategies, and synchronize the early warning information to the soil environmental monitoring platform. The data storage unit is electrically connected to both the data acquisition unit and the intelligent control unit and is configured to adopt a cloud storage architecture to store raw testing data, standardized data, data management reports, soil quality level results, early warning information, and soil testing-specific knowledge graph data throughout their entire lifecycle. It supports data traceability, retrieval, and backup, and the data storage unit is equipped with an access permission management module to enable graded data access for staff at different levels.

[0041] Specifically, the early warning push unit and the intelligent control unit work together to push early warning information and improvement and detection strategies in a tiered manner, synchronizing them to the regulatory platform. This establishes a seamless data management and environmental supervision link, improving the timeliness and coordination of soil pollution prevention and control. The data storage unit adopts a cloud storage architecture to achieve full lifecycle storage of all types of soil testing data, supporting traceability, retrieval, and backup to ensure data integrity and traceability. An access permission management module is configured to enable tiered data access, accommodating the business data needs of staff at different levels while mitigating the risks of data leakage and tampering, ensuring data management security.

[0042] Example 1: This embodiment discloses a specific implementation of a big data-based intelligent management system for soil testing and inspection data. Taking the intelligent management of soil testing data within a 10km² area surrounding a provincial-level chemical industrial park as a practical application scenario, the composition of the system, the execution flow of each unit's functions, and the overall collaborative operation logic are described in detail. The system includes a data acquisition unit, a data preprocessing unit, a model analysis unit, an intelligent control unit, an early warning push unit, and a data storage unit. Each unit is electrically connected via an industrial Ethernet network, enabling bidirectional interaction between data and commands. This forms a closed-loop intelligent management system for soil testing and inspection data, from acquisition, processing, and analysis to control, early warning, and storage. The overall system architecture is compatible with industry standards such as the "Soil Environmental Quality Standard for Construction Land Soil Pollution Risk Control (GB 36600-2018)" and the "Soil Environmental Quality Standard for Agricultural Land Soil Pollution Risk Control (GB 15618-2018)," enabling standardized, intelligent, and refined management of soil testing data. The implementation process of each unit is described in detail below.

[0043] I. Implementation and Operation of the Data Acquisition Unit In this embodiment, the data acquisition unit adopts a dual-mode acquisition method of mobile detection terminal + fixed monitoring station, and is configured to collect multi-dimensional data of soil testing and detection in the area to be managed around the industrial park. The specific acquisition dimensions include five categories: soil physicochemical index data, pollutant detection data, sampling geospatial information data, detection equipment operation data, and historical soil testing and detection data of the area over the past 5 years. Among them, physicochemical index data includes pH value, organic matter content, soil bulk density, porosity, etc.; pollutant detection data focuses on characteristic pollutants of chemical industrial parks, including heavy metals (Pb, Cd, Cr, Hg), volatile organic compounds (VOCs), semi-volatile organic compounds (SVOCs), etc.; sampling geospatial information data is obtained through GPS positioning module with an accuracy of 1m; detection equipment operation data includes portable soil tester serial number, calibration time, detection range, and operating status; historical testing and detection data is retrieved from the local ecological environment bureau's soil environmental monitoring database, and a total of 12,000 valid historical data entries are obtained.

[0044] The data acquisition unit sets up 50 sampling points (numbered A1-A50) in the area to be managed using a grid-based method, with sampling depths of 0-0.5m (top soil) and 0.5-1.5m (deep soil). After data acquisition, the multi-dimensional data is classified into structured data (such as pH value and heavy metal concentration) and unstructured data (such as equipment operation logs and sampling site images). A raw soil testing database is built based on the HDFS distributed storage architecture, and four-dimensional basic labels are added to each data entry, including sampling time, sampling point, testing item, and sampling depth. For example, the label is "2025-5-22|A12|Heavy metal Cd|0-0.5m", which realizes the initial classification and identification of the data.

[0045] II. Implementation of the Data Preprocessing Unit The data preprocessing unit is electrically connected to the data acquisition unit. After receiving all the data from the original database, it completes the data preprocessing according to the steps of data cleaning → missing value completion → unit unification and normalization. The specific implementation is as follows: 1. Data Cleaning: The preprocessing unit has built-in invalid data identification rules, which automatically remove invalid data caused by sampling operation errors (such as inconsistent sampling depth), detection equipment failure (such as uncalibrated equipment), and human recording errors (such as values ​​exceeding the reasonable range). In this embodiment, a total of 186 invalid data entries were removed, and the original data validity rate was increased from the initial 98.45% to 99.85%. 2. Missing Value Completion: For missing values ​​in soil testing indicators, the **mean interpolation method based on soil types and areas within the same region** is used. For example, if the Cd concentration data for the 0-0.5m soil layer at point A8 is missing, the Cd concentration values ​​(0.30 mg / kg, 0.35 mg / kg, and 0.31 mg / kg) of the same soil layer at adjacent points A7, A9, and A10 are retrieved, and the mean value (0.32 mg / kg) is calculated to complete the interpolation. For missing values ​​related to geographic space, sampling time, etc., the source verification and completion method is used, linking the sampling personnel's mobile terminal sampling record sheets to complete the information. In this example, a total of 62 missing data entries were completed. 3. Dimensional Unification and Normalization: All completed numerical data are normalized to eliminate dimensional differences using the following preprocessing formula: in, For standardized data, This is the original test data. It is the minimum value of the test data for the same project. This represents the maximum value of the test data for the same project.

[0046] In this embodiment, the original measured values ​​of Pb concentration in the soil of the area to be managed ranged from 0.15 mg / kg to 1.20 mg / kg. The original Pb concentration at point A15 was 0.60 mg / kg, and its standardized value was calculated to be: After the preprocessing unit completes the full data processing, it outputs standardized soil test data, which is then synchronized to the model analysis unit.

[0047] III. Implementation and Operation of the Model Analysis Unit The model analysis unit is electrically connected to the data preprocessing unit. After receiving standardized soil testing data, it completes three core operations: intelligent management model construction and training, soil testing-specific knowledge graph construction, and data anomaly identification. The specific implementation is as follows: 1. Intelligent Management Model Construction and Training: The model analysis unit is configured to integrate the random forest machine learning algorithm with soil testing and inspection industry rules (such as pollutant threshold standards in GB 36600-2018 and GB 15618-2018) to construct a multi-dimensional data association intelligent management model. The model input is standardized soil testing data, and the output is data association, indicator change trends, data quality rating, and anomaly identification results. The model is initially trained using 12,000 historical testing data points from the area to be managed over the past 5 years, with 500 training iterations, and the model fit R² reaches 0.92. Simultaneously, incremental training is performed using 50 testing data points collected in real time per batch, dynamically updating the model parameters so that the accuracy of anomaly identification and indicator analysis continuously improves with data accumulation. In this embodiment, after incremental training, the anomaly identification accuracy reaches 91.5%.

[0048] 2. Construction and Update of Soil Testing-Specific Knowledge Graph: The model analysis unit constructs a soil testing-specific knowledge graph based on the graph database Neo4j. Information such as sampling points, soil type (primarily alluvial soil in this area), testing indicators, pollutant types, pollution levels, soil remediation measures, and historical testing data are used as graph nodes, and the relationships between nodes are used as graph edges, forming a visualized knowledge network. For example, the node association "A25 point → alluvial soil → Cd concentration 0.45 mg / kg → mild heavy metal pollution → application of phosphate rock passivating agent" is constructed. Simultaneously, the model analysis unit dynamically updates the knowledge graph based on real-time collected soil testing data. After adding new sampling points and testing indicator data, the graph nodes and edges are automatically expanded, maintaining the timeliness of the graph information.

[0049] 3. Data Anomaly Identification: The model analysis unit calculates the correlation value between real-time detection data and historical data in the same area, as well as the analysis residual value of the real-time detection data, using the trained intelligent management model. In this embodiment, the preset correlation threshold for soil detection data is 0.7, and the residual threshold is 0.2. The calculated value of the real-time data is compared with the threshold to determine whether the data is abnormal, and the anomaly type is identified as sampling anomaly, detection anomaly, data transmission anomaly, or indicator mutation anomaly. For example, the correlation value of the real-time Cd concentration detection data at point A30 is 0.65, and the analysis residual value is 0.23, both exceeding the threshold. The model analysis unit determines that the data is abnormal and identifies the anomaly type as detection anomaly (the portable detector has not completed the daily calibration) through equipment operation data tracing.

[0050] The model analysis unit synchronizes the anomaly identification results, knowledge graph data, and indicator trend analysis results to the intelligent control unit.

[0051] IV. Implementation and Operation of Intelligent Control Unit The intelligent control unit is electrically connected to the model analysis unit. After receiving the above data, it performs operations according to two main branches: handling abnormal data and managing non-abnormal data. Finally, it completes the comprehensive soil quality score, quality grade determination, and data management report generation. The specific implementation is as follows: 1. Targeted handling of abnormal data: Based on the anomaly type identification results from the model analysis unit, the intelligent control unit performs corresponding correction and completion operations: For sampling / detection anomalies (such as detection anomalies at point A30), a retest instruction for the corresponding sampling point is automatically triggered and pushed to the mobile terminal of the testing personnel. After the retest is completed, the valid test data of 0.40 mg / kg will replace the abnormal data and be updated to the original database and knowledge graph simultaneously. In case of data transmission anomalies, automatically perform data retransmission and MD5 verification operations until the data transmission is error-free; For any abnormal sudden changes in indicators, the data is first marked, and then the data from three sampling points in the same area are linked together for verification to confirm whether it is an actual change in soil indicators. If it is an actual change, it is synchronized to the knowledge graph and included in the trend analysis.

[0052] 2. Intelligent management of non-abnormal data: Multi-level classification and archiving: Non-abnormal data is classified and archived in three levels according to soil use type (construction land / agricultural land), testing items (heavy metals / VOCs / SVOCs), and pollutant type. A data association retrieval system is established to support precise retrieval based on multiple conditions such as "sampling point + testing item + sampling time". For example, searching for "A10 | heavy metal Cd | May 2025" can retrieve the full information of the data within 1 second. Time-series trend analysis: Based on big data time-series analysis technology, the heavy metal concentration detection data of points A1-A50 for 6 months are analyzed to obtain the index change pattern. In this example, the monthly average change rate of Pb concentration is 0.02 mg / kg, the change cycle is 3 months, and it is positively correlated with the emission intensity of chemical production in the park. A visualized index change trend line chart is generated. Knowledge graph linkage application: Retrieve the soil testing-specific knowledge graph constructed by the model analysis unit to achieve intelligent knowledge retrieval of the graph, and perform intelligent reasoning based on the association rules between real-time detection data and the graph; for example, searching for "Cd pollution of tidal soil around chemical industrial park" can quickly obtain corresponding soil improvement measures; for the A25 point with a Cd concentration of 0.45mg / kg (slight pollution), reasoning can derive optimization suggestions for the detection point (increase sampling by 1 time per month) and targeted improvement suggestions (apply phosphate rock passivating agent), providing knowledge-level decision support for soil environmental remediation.

[0053] 3. Comprehensive soil quality scoring and grading: The intelligent control unit calculates a comprehensive soil quality score based on standardized data from various soil testing indicators, combined with industry standard thresholds, using the following scoring formula: Among them, among them, For the comprehensive soil quality score, To achieve the target compliance rate, For data integrity, To assess the health of the indicator trend, The weighting coefficient is used; in this embodiment, the weighting coefficient is preset. =0.5 (Indicator compliance rate, core weight) =0.2 (data integrity) =0.3 (indicator trend health), and None of them are 0.

[0054] Taking the cluster of points A1-A10 in the area to be managed as an example, the indicator compliance rate Data integrity Health of indicator trends The overall score calculated is as follows: S=0.5×0.85+0.2×0.98+0.3×0.70=0.425+0.196+0.21=0.831 The intelligent control unit is pre-configured with three preset scores for soil quality, where the first preset score < the second preset score < the third preset score. In this embodiment, the first preset score is set to 0.6, the second preset score to 0.8, and the third preset score to 0.9. The soil quality level is determined according to the score range. Overall score > 0.9: Excellent; 0.8 < Overall score ≤ 0.9: Good; 0.6 < Overall score ≤ 0.8: Medium; Overall score ≤ 0.6: Poor.

[0055] The cluster of points A1-A10 received a comprehensive score of 0.831, which is considered good; while the cluster of points A28-A30 received a comprehensive score of 0.58, which is considered poor. The intelligent management unit, combined with the reasoning results from the knowledge graph, generated targeted soil improvement suggestions for the area.

[0056] 4. Data Management Report Generation: Based on the above analysis results, the intelligent control unit automatically generates a soil testing data management report. The report includes core content such as data quality assessment of the area to be managed (effective data rate of 99.85%), analysis of changes in testing indicators, data classification and archiving status, recommendations for key testing points, and soil improvement plans. The report is synchronized to the data storage unit and the early warning push unit.

[0057] V. Implementation and Operation of the Early Warning Push Unit The early warning push unit is electrically connected to the intelligent control unit and is configured to receive the soil quality grade determination results and graded early warning prompts from the intelligent control unit. It then pushes early warning information and synchronizes it with the monitoring platform according to a two-level early warning mechanism. In this embodiment, the following is preset: Level 1 warning: Soil quality grade is poor (overall score ≤ 0.6), and the information is sent to the environmental management department of the industrial park and the local municipal ecological and environmental bureau. Level II warning: Soil quality grade is medium (0.6 < comprehensive score ≤ 0.8), and the information is sent to the park's soil testing station and regional environmental monitoring points.

[0058] The early warning push unit pushes early warning information through a dual channel of government intranet and mobile terminal APP. The information includes the location of the polluted area, the type and concentration of pollutants, the degree of pollution, targeted improvement suggestions, and the strategy for adjusting the frequency of testing. For example, for the first-level early warning of the A28-A30 cluster, the push content is "The average Cd concentration of the soil at the A28-A30 site (southeast side of the chemical industrial park) is 0.48 mg / kg, the quality level is poor, and it is recommended to apply phosphate rock passivating agent and conduct sampling and testing twice a month". At the same time, all early warning information is synchronized to the provincial soil environmental supervision platform to achieve data exchange with the higher-level regulatory departments.

[0059] VI. Implementation and Operation of Data Storage Units The data storage unit is electrically connected to the data acquisition unit and the intelligent control unit. It adopts Alibaba Cloud's distributed cloud storage architecture and is configured to store soil testing raw data, standardized data, data management reports, soil quality grade results, early warning information, and soil testing-specific knowledge graph data throughout their entire lifecycle. It supports data traceability, retrieval, and automatic backup.

[0060] In this embodiment, the data storage unit is equipped with a three-level access control module to enable hierarchical data access for staff at different levels. The access control configuration is as follows: 1. Level 1 privileges (System Administrator): Can access and edit all data, responsible for system parameter updates and maintenance; 2. Level 2 Access (Testing Personnel): Can access and view raw and standardized soil testing data for this area, and is responsible for data entry and retesting operations; 3. Level 3 Authority (Environmental Regulatory Personnel): Can access and view soil quality grade results, early warning information, and data management reports, and is responsible for environmental supervision and governance decisions.

[0061] Meanwhile, the data storage unit is equipped with an automatic backup mechanism every morning at midnight, backing up data to a remote server to prevent data loss. It also supports full-process data traceability, allowing users to query the entire process information of sampling, testing, preprocessing, and analysis of any test data, thus meeting the traceability management needs of soil test data.

[0062] VII. System Overall Operation Closed Loop The overall operation process of this big data-based intelligent management system for soil testing and inspection data is as follows: the data acquisition unit collects multi-dimensional data and builds a raw database → the data preprocessing unit completes data standardization processing → the model analysis unit builds and trains an intelligent management model, constructs a knowledge graph, and identifies abnormal data → the intelligent control unit handles abnormal data, manages non-abnormal data, and completes soil quality scoring and rating → the early warning push unit pushes early warning information according to the level and synchronizes it to the monitoring platform → the data storage unit realizes full lifecycle storage and hierarchical access of all data. All units work together to form an intelligent management closed loop for soil testing and inspection data.

[0063] VIII. Implementation Results After its application in the management of soil testing data around the provincial chemical industrial park, this system achieved the following technical effects: 1. The effectiveness of soil testing data has increased from 82% in the traditional management model to 99.85%, and the automation level of data preprocessing has reached 100%, significantly reducing manual operation costs; 2. Data retrieval efficiency is improved by more than 70%, and the response time for multi-condition accurate retrieval is ≤1 second, solving the pain points of fragmentation and difficulty in retrieval in traditional data management; 3. The accuracy rate of soil pollution trend prediction reached 89%. Combined with knowledge graphs, it achieved an upgrade from "data management" to "knowledge empowerment", providing accurate decision support for soil environmental governance. 4. Data interoperability with provincial and municipal soil environmental monitoring platforms has been achieved, with early warning information push response time ≤5min, improving the timeliness and effectiveness of soil pollution prevention and control.

[0064] This system can be widely used in soil testing and inspection data management in various areas such as industrial parks, farmland, and mining areas. It is adaptable to the needs of different soil types and testing projects. Through the deep integration of big data technology and the soil testing industry, it effectively improves the standardization, intelligence, and precision of soil testing data management.

[0065] Those skilled in the art should understand that this embodiment is merely an illustrative example and is not intended to limit the system. In practical applications, system parameters (such as weighting coefficients, preset scores, and early warning thresholds) can be adjusted according to the soil characteristics and detection requirements of the area to be managed. The functional modules of each unit can be flexibly expanded according to the actual scenario, without departing from the core protection scope of this invention.

[0066] 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.

[0067] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A big data-based intelligent management system for soil testing and inspection data, characterized in that: It includes a data acquisition unit, a data preprocessing unit, a model analysis unit, and an intelligent control unit; The data acquisition unit is configured to collect multi-dimensional soil testing data from the area to be managed and construct a raw soil testing database. The data preprocessing unit is electrically connected to the data acquisition unit and is configured to preprocess the data in the original database, build an intelligent management model for soil testing data based on big data technology, and input the preprocessed data into the management model. The model analysis unit is electrically connected to the data preprocessing unit and is configured to determine whether the real-time collected soil testing data is abnormal based on the output of the management model. The intelligent control unit is electrically connected to the model analysis unit and is configured to perform corresponding correction and completion operations on the soil testing data according to the anomaly type when the data is determined to be abnormal; and to perform intelligent classification, archiving and trend analysis on the soil testing data when the data is determined to be non-abnormal, and to determine the soil quality level of the area to be managed in conjunction with soil testing industry standards, while generating a data management report.

2. The intelligent management system for soil testing and inspection data based on big data as described in claim 1, characterized in that, The data acquisition unit is configured to acquire physicochemical index data, pollutant detection data, sampling geospatial information data, detection equipment operation data, and historical inspection and testing data of the soil in the area to be managed. It is also configured to classify the multi-dimensional data into structured and unstructured data, build a soil detection raw database based on a distributed storage architecture, and add basic tags such as sampling time, sampling point, and detection items to each data entry.

3. The intelligent management system for soil testing and inspection data based on big data as described in claim 2, characterized in that, The data preprocessing unit is configured to perform data cleaning operations on the raw data to remove invalid data caused by sampling operation errors, detection equipment failures, and manual recording errors. It is also configured to complete the missing values ​​of the cleaned data, using the mean interpolation method of soil of the same type in the same region for soil testing professional indicators, and the source verification and completion method for geographic information and time information; Simultaneously, it is configured to perform dimensional unification and normalization on the completed data to obtain standardized soil testing data. The preprocessing formula is: in, For standardized data, This is the original test data. It is the minimum value of the test data for the same project. This represents the maximum value of the test data for the same project.

4. The intelligent management system for soil testing and inspection data based on big data as described in claim 1, characterized in that, The model analysis unit is configured to integrate machine learning algorithms with soil testing and inspection industry rules to construct a multi-dimensional data association intelligent management model. The model input is standardized soil testing data, and the model output includes data association relationships, indicator change trends, data quality ratings, and anomaly identification results. It is also configured to use historical soil testing data of the area to be managed to initially train the management model, and combine it with real-time collected soil testing data to perform incremental training, dynamically update the model parameters, and improve the model's recognition and analysis accuracy. It is also configured to build a dedicated knowledge graph for soil testing, which will perform multi-dimensional correlation modeling of sampling geospatial information, soil type, testing indicators, pollutant type, pollution level, soil improvement measures and historical testing data to form a visual knowledge association network. The knowledge graph will be dynamically updated in conjunction with real-time collected soil testing data to maintain the timeliness of the graph information.

5. The intelligent management system for soil testing and inspection data based on big data as described in claim 4, characterized in that, The model analysis unit is configured to calculate the correlation value between real-time detection data and historical data in the same area, as well as the analysis residual value of real-time detection data, through the management model. It is also configured to preset correlation thresholds and residual thresholds for soil testing data, compare the calculated values ​​of real-time data with the thresholds to determine whether the data is abnormal, and identify the type of abnormality as sampling abnormality, detection abnormality, data transmission abnormality, or indicator mutation abnormality.

6. The intelligent management system for soil testing and inspection data based on big data as described in claim 5, characterized in that, The intelligent control unit is configured to, when the anomaly type is identified as sampling anomaly or detection anomaly, trigger a supplementary sampling instruction for the corresponding sampling point or a re-inspection instruction for the detection item, replace the abnormal data with the valid data after supplementary sampling and re-inspection, and update the original database. It is configured to perform data retransmission and verification operations when the anomaly type is identified as a data transmission anomaly, until the data transmission is error-free; it is configured to mark the data and correlate it with data from surrounding points in the same area when the anomaly type is identified as an indicator mutation anomaly, to verify whether it is a change in actual soil indicators.

7. The intelligent management system for soil testing and inspection data based on big data as described in claim 1, characterized in that, The intelligent control unit is configured to classify and archive non-abnormal data in multiple levels according to soil use type, testing items, and pollutant type, establish a data association retrieval system, and support accurate retrieval based on multiple conditions such as sampling point, time, and indicators. It is also configured to perform trend analysis on soil testing indicators at each sampling point based on big data time series analysis technology, obtain the rate of change, cycle of change and correlation of the indicators, and generate a visualized trend chart of indicator changes. It is also configured to retrieve the soil testing-specific knowledge graph constructed by the model analysis unit, realize intelligent knowledge retrieval of the graph, and perform intelligent reasoning based on the association rules between real-time detection data and the graph, providing knowledge-level decision support for optimizing soil testing points in the area to be managed, formulating targeted soil improvement plans, and predicting pollution trends.

8. The intelligent management system for soil testing and inspection data based on big data as described in claim 7, characterized in that, The intelligent control unit is configured to calculate a comprehensive soil quality score based on standardized data from various soil testing indicators, combined with industry standard thresholds. The scoring formula is as follows: in, For the comprehensive soil quality score, To achieve the target compliance rate, For data integrity, To assess the health of the indicator trend, These are the weighting coefficients, and None of them are 0; It is also configured to generate a data management report based on the comprehensive score. The report includes a quality assessment of soil testing data in the area to be managed, analysis of index changes, data archiving status, and recommendations for subsequent testing.

9. The intelligent management system for soil testing and inspection data based on big data as described in claim 8, characterized in that, The intelligent control unit is configured to pre-configure a first preset score, a second preset score, and a third preset score for soil quality, wherein the first preset score < the second preset score < the third preset score; The system is configured to determine the soil quality grade as excellent when the comprehensive soil quality score is greater than the third preset score, good when the comprehensive score is greater than the second preset score and less than or equal to the third preset score, medium when the comprehensive score is greater than the first preset score and less than or equal to the second preset score, and poor when the comprehensive score is less than or equal to the first preset score. It is also configured to generate targeted soil improvement suggestions based on the type of pollutants and the trend of indicator changes, combined with the reasoning results of knowledge graph, when the soil quality level is medium or poor, and trigger graded early warning prompts.

10. The intelligent management system for soil testing and inspection data based on big data according to claim 1, characterized in that, It also includes an early warning push unit and a data storage unit. The early warning push unit is electrically connected to the intelligent control unit and is configured to receive graded early warning prompts when the soil quality level is medium or poor, push soil improvement suggestions and detection adjustment strategies, and synchronize the early warning information to the soil environmental monitoring platform. The data storage unit is electrically connected to the data acquisition unit and the intelligent control unit. It is configured to adopt a cloud storage architecture to store raw test data, standardized data, data management reports, soil quality grade results, early warning information and soil test-specific knowledge graph data throughout their entire lifecycle. It supports data traceability, retrieval and backup. The data storage unit is also equipped with an access permission management module to enable hierarchical data access for staff at different levels.

Images